Potential Future Directions in Optimization of Students' Performance Prediction System

Previous studies widely report the optimization of performance predictions to highlight at-risk students and advance the achievement of excellent students. They also have contributions that overlap different fields of research. On the one hand, they have insightful psychological studies, data mining discoveries, and data analysis findings. On the other hand, they produce a variety of performance prediction approaches to assess students' performance during cognitive tasks. However, the synchronization between these studies is still a black box that increases prediction systems' dependency on real-world datasets. It also delays the mathematical modeling of students' emotional attributes. This review paper performs an insightful analysis and thorough literature-based survey to draw a comprehensive picture of potential challenges and prior contributions. The review consists of 1497 publications from 1990 to 2022 (32 years), which reported various opportunities for future performance prediction researchers. First, it evaluates psychological studies, data analysis results, and data mining findings to provide a general picture of the statistical association among students' performance and various influential factors. Second, it critically evaluates new students' performance prediction techniques, modifications in existing techniques, and comprehensive studies based on the comparative analysis. Lastly, future directions and potential pilot projects based on the assumption-based dataset are highlighted to optimize the existing performance prediction systems.


Introduction
Over the past few decades, students' performance has been predicted while evaluating the influence of different factors, such as emotional attributes, family attributes, study schedule, institutional attributes, and students' scores in assignments, quizzes, and final examinations [1][2][3][4][5]. Such systems provide useful applications to a wide area in academia, i.e., students' success and failure estimation due to influential factors [6][7][8][9][10]. is study splits the earlier contributions into two groups. e first group consists of insightful psychological studies, data mining discoveries, and data analysis findings that indirectly contribute to the optimization of students' performance prediction systems. e second group reports the optimization of existing prediction systems based on the findings of the first group. However, the extensive synchronization between the two groups is still a black box that ultimately increases students' performance prediction systems' dependency on a real-world dataset. Such synchronization can provide useful ideas during the optimization and data collection process. It also paves the way for an assumption-based dataset to prove the viability of pilot project implementations that will speed up modeling students' emotional attributes.
is review paper conducts an insightful study and literature-based survey to draw a comprehensive picture of the prior studies on student performance analysis and prediction. e review consists of 1497 articles from 1990 to 2022 (32 years), which reported various information for future researchers: 1. It explores and lists the research fields' contributions focusing on students' performance optimization. Psychology and data analysis fields pave the way for effective solutions to the problems of data deficiency. ey provide qualitative findings that can be used for creating an assumption-based dataset for pilot project implementation. 2. It thoroughly considered new and modified algorithms that predict students' performance. Also, a comparative analysis was performed between the existing students' performance prediction approaches to provide better recommendations for optimization. 3. e study delivers a comprehensive picture of potential challenges and research direction for future researchers. e review also shows that very few contributions have mathematically modeled emotional attributes. e remaining sections of this review are as follows: Section 2 gives a detailed literature review. Section 3 elaborates the review methodology, and Section 4 produces data evaluation. Section 5 presents future challenges, and Section 6 concludes the study.

Contributions of Psychology.
Psychological studies results manifest that students' performance is easily influenced via emotional attributes, such as frustration, anxiety, stress, over expectation of parents, and parents' relationship [35][36][37]. e results provide correlations statistics among emotional factors and the expected performance of students in cognitive activity, such as attempting the examination, quizzes, assignments, class activities, and extracurricular activities. ese particular emotional factors can negatively and positively impact the students' performance. In such a situation, emotional severity, family attributes, and institutional factors play a crucial role in influencing performance [38][39][40]. It shows that performance is always very sensitive and affected by the individuals' surroundings.

Contributions of Data Mining.
e data mining evaluates the relationship among various students' factors, such as the role of emotional factors, family attributes, institutional factors, and class performance. Such studies provide good opportunities for accurate estimations of expected students' performance [41][42][43][44][45][46][47]. e meaningful patterns always produce good directions for further exploration. e previous articles lack coordination between the students' influential attributes and academic performance. e literature lacks accurate techniques to simulate students' performance due to the insufficient synchronization and coordination among earlier studies on students' factors. e mathematical modulation of students' performance needs to formulate the function of student factors. However, it is inspiring to closely examine the quantitative influence of several student factors on academic achievements. e earlier studies show that emotional, family, study schedule, and institutional attributes are the significant factors that can easily influence students' academic performance in any critical cognitive activity. Prior studies illustrate that educational data mining practices contribute to students' factor evaluation process and performance prediction. Institutional factors involve teaching methodology, engaging students in the classroom, and the vision of instruction. According to literature studies, teachers play an active role in institutional attributes influencing students' performance. ey provide administrative assistance and assistance in ensuring discipline [48][49][50][51][52][53][54][55][56][57].

Synchronization among Existing Studies.
Accurate performance prediction needs to examine students' factors beyond the computer science framework. e literature studies are still limited in finding an authentic and extendable approach that overlaps psychology, data mining, data analysis, and cognitive research. Articles have various solutions to predict students' performance using different techniques that could have the potential to be escalated to more general problems of predicting student performance [58][59][60][61][62]. e primary objective of the current review attempt is to efficiently explore the relationship between students' factors (as mentioned earlier) and their performance. erefore, the literature is studied with the selected students' attributes (emotional, family, study schedule, and institutional) and effects. Articles show that most students do not participate in extracurricular activities, believing extra activities would negatively affect their academic achievements. Earlier studies also focus on predicting college students' performance by considering all the important aspects. ey delivered a prediction system to estimate performance by assisting the university in selecting each candidate using past academic records of students granted admissions [29,[63][64][65][66][67][68][69]. Such efforts show that earlier studies contribute to decreasing the number of at-risk students and advancing the performance of excellent students.
Literature also attempted to perform a survey on classroom learning in different environments. It analyzes various aspects and factors influencing (positively or negatively) performance in a classroom that interfere with learning.
is paper presents a systematic review of numerous studies on students' performance in classroom learning. For a few decades, the research has produced numerous results in students' performance evaluation; however, the education system needs a complete and detailed performance prediction system that can ensure interaction and coordination between the aforementioned students' factors. Literature studies delivered various contributions, such as the proposal of an innovative model that targets modifying learning sustainability through smart education applications and regression and correlation among students' factors, and logistic regression analyses generated that being female, first-semester GPA, number of courses per regular semester, and number of courses per summer semester were imperative predictors of baccalaureate degree achievement [70][71][72][73][74][75][76][77].

Existing Models and Performance Prediction System
Optimization. Studies have focused on applying artificial neural networks to predict performance in different environments. Articles are also saturated with deep learning techniques that deliver prediction and highlight at-risk students. Few other technologies provide opportunities to accurately evaluate the performance and reduce the failure rates [78][79][80][81][82]. It also helps in counseling students in alarming situations that can positively impact their academic achievements, i.e., COVID-19. us, during the literature survey, we have found many students' prediction systems which are interesting; nevertheless, they are failed to mathematical model emotional attributes and synchronized them with institutional attributes, study schedules, and family attributes [31,37,[83][84][85][86][87][88][89][90][91][92][93][94][95][96][97][98]. e objective of this study is to identify the relationship between extracurricular activities and students' performances. e articles deliver many results on the effects of influential students' factors. is study explores performance prediction beyond the scope of computer science and machine learning.

Related Performance Prediction
Methods. As discussed earlier, many studies solved meaningful challenges in students' performance prediction area of research for a few decades. e earlier studies have many contributions in the form of neural works, recommendation systems, course recommendations, and students' performance evaluation systems [87,[99][100][101][102][103][104][105][106][107][108][109]. e prior studies demonstrate comprehensive work on students' performance prediction systems that use information obtained during the interaction of students with the institutional attributes. To mathematically consider the expected actions of a students' factors, such information provides proper guidelines. e significant characteristic is the identical structure of the information processing system of students, which can be replicated to construct a learning algorithm (cognitive architecture). Literature studies are flooded with many findings that primarily contribute to prediction algorithms and mathematical models; nevertheless, modeling the relationship between students' emotions (frustration, stress, etc.) and students' performance is very little focused [110][111][112][113][114][115].
Also, the published studies on modeling emotion are not extendable toward a matured prediction system. So, the dire need is to assess the main framework of existing prediction algorithms. Exploring the qualitative results of psychological studies and data analysis discoveries is needed to estimate students' performance. It will also help in the iterative calculation of emotional influence on performance during critical cognitive activities. Extraordinary academic performance is only possible with excellent cognitive skills. Such skills are needed to accomplish any task requiring problem-solving approaches, reasoning, and memory management. However, with inadequate cognitive abilities, an individual cannot achieve an excellent score in various cognitive tasks, i.e., assignments, quizzes, and written examinations. ey require students to process new information, organize learning, and retrieve that data (from memory) for later use. So, predicting performance while calculating the intense impact of various groups of factors is crucial not only for tutors to ensure effective teaching methodology but also for students' achievements and effective academic policies. Earlier studies have delivered many approaches that predict students' performance; nevertheless, they have paved the way for new challenges for effective educational systems. e skills levels of students are changing as they learn and forget. e educational system needs such a system that can manage the students' dynamic behavior during cognitive activities.
Other studies have described the students' personality traits and the essential characteristics of personality. Results reveal that performance prediction design can be broken down into subsections that are more realistic in comparison to other techniques. It also paves the way for the development of performance prediction architectures which were easy to understand. e current review illustrates that the performance prediction system needs to coordinate among prediction architecture, psychological experiments, semantical investigations, statistical analysis, and mathematical formulation.
is work provided an outstanding opportunity for researchers belonging to performance prediction, bioinformatics, data mining, and data integration.

Review Methodology
e review process is started from the initial screening within the scope of the current attempt. As elaborated earlier, this review focused on recent and state-of-the-art contributions to student performance prediction. Figure 1 illustrates the methodology of the review. We have divided the complete review process into the following sections.

Review Process.
is study reviews the earlier studies thoroughly based on the procedures prescribed by Petersen et al. [116] and Keele [117]. e methodology is adopted from Keele while the study mapping method is copied from Petersen et al. e review process is initiated with the Computational Intelligence and Neuroscience modified procedure, which is demonstrated in Figure 1. For better understanding, the review delivers a detailed methodology of the prior work contributing to student performance prediction directly or indirectly. Moreover, the study put a list of research questions to demonstrate the main objectives. ese research questions enable us to choose relevant research studies for screening and investigating the main challenges in students' performance predictions. Every research question has a list of keywords to explore the literature and learn about a particular question. ese keywords are used to search publications, including peerreviewed book chapters, conferences proceeding, and journal articles.   disadvantages [116]. (1) e automatic search is not feasible for the current review [118]. (2) e manual searching strategy gives more relevant studies. Table 1 reflects the list of keywords that have produced a variety of articles published by various publishers, i.e., IEEE, Elsevier, Springer, Hindawi, MDPI, ACM, Wiley, and others. It shows many articles, including journals, book chapters, and conference proceedings.

Research Questions
e keywords were searched directly on publishers' websites and Google Scholar with a default setting. We have evaluated all the articles and collected those that deliver relevant findings for further screening. Furthermore, the main factors, topics, and relevant studies, including journals, conference proceedings, and book chapters, are given below: 1. Emotional attributes 2. Family factors 3. Study Schedule 4. Institutional attributes 5. Psychology, data mining, and data analysis findings on the factors as mentioned earlier 6. Contribution of cognitive computing, deep learning, and machine learning in students' performance prediction 7. Reviews and comparison 3.4. Screening. Screening of studies is performed with the following terms and conditions: 1. e team selected the publications of the more relevant journal, conference, and book chapter. 2. Second, we have focused on the relevant title with impressive citations in Google Scholar. 3. ird, rapid reviews were performed for further evaluation and data extraction. During the rapid review, we have focused on the abstract and introduction to get some idea about the challenges, motivations, and contributions. ese three steps were performed to create a database for further information extraction and data collection.

Information Collection.
Various information was extracted from the selected publications during the information collection process, which are shown in Table 1 to 4 Also, a spreadsheet was used to record the various information for further consideration of the research questions.
e recorded data are shown in the tables mentioned above.

Q1: What Are the Applications of Student Performance
Prediction Systems. A performance prediction system is essential to predict at-risk students to devise a solution for successful graduation and goal achievements, such as special treatment and counseling sessions. Such prediction systems are more challenging due to the significant factors affecting students' performance. us, a systematic review of the literature has been performed to highlight potential issues in predicting student performance. e study also shows the contributions of previous articles beyond the scope of artificial intelligence, i.e., data mining, data analysis, and psychology techniques contributing to performance prediction. Also, this study provides an overview of prediction techniques that have been used to estimate performance. It focuses on how the predictive algorithm can be used to identify key attributes in influencing students' academic achievements. With the help of data mining and machine learning techniques in education, the study could have a more effective methodology in proposing a new prediction algorithm and modifying existing students' performance prediction systems. e primary application outcomes of students' performance prediction are given below.

Prediction of At-Risk Student.
It is crucial to predict atrisk students and devise an effective learning environment in classrooms and laboratories. Although the literature studies are saturated with tremendous results, it is still challenging as the prediction system cannot synchronize and mathematically model emotional attributes, family issues, study schedules, and institutional attributes to develop a significant prediction system. e current review's first target is to highlight the possibilities of predicting at-risk students while coordinating between literature studies.

Advances the Students' Academic Achievements.
e performance prediction system is essential for at-risk, average, and excellent students. e influential factors that drive academic achievement are an eternal global challenge associated with students, families, teachers, and educational policymakers. Exploring these factors benefits all those interested in developing a system for students' performance prediction worldwide. Suppose the prediction system considers a large number of influential factors. In that case, the academic achievement of excellent students can also be advanced, i.e., the prediction system could highlight problems due to various emotional, study schedules, family, and institute-related attributes.

Monitoring Students'
Behavior. Student behavior plays a significant role in improving academic achievements, such as interaction and attitude with the teachers, seriousness, and unseriousness in the classroom. Articles of psychology and data analysis contribute to student behavior evaluation, merits, and demerits of various aspects of behavior. We need a prediction system that efficiently modulates the relationship between behavior and students' performance to highlight, monitor, and improve students' interaction and engagement in the classroom. It is also essential for the institution to devise effective controlling policies to counter and control the demerit of various behaviors. rough such a prediction system, teachers can easily guide their students in setting and achieving academic goals. A teacher can also help students understand their behavior and its impact on Computational Intelligence and Neuroscience others. e adverse effects of behavior can be overcome and later on monitored by supervising students. Such a system enhances the overall reputation of the institution. Other benefits include preventing early school drop-ups and building good relationships among students. According to Kennelly and Monrad [119,120], the behavioral problem plays a key role in indicating students at risk and highlighting the individuals near to being dropped off at the institute. erefore, employing strategies to monitor and control student behavior is extremely important for an effective educational system in a society.

Q2: What Are the Factors at Can Optimize Student Performance Prediction?
e literature studies indicate that many factors influence students' performance in cognitive activities, such as quizzes, assignments, examinations, and homework. It includes family-related factors, emotional factors, gender description, and institution-related factors. A brief description of these factors is given below.

Family-Related Factors.
e parental involvement and their particular influence are two-fold. First, the earlier studies claim that the interaction of parents positively influences performance. It enhances the academic achievements of the student in critical environments. Research results highlight that parents' friendly attitudes positively affect student performance, such as daily engagement in cognitive activities. Positive parent involvement can advance the performance, and that father or mother is the first teacher who plays the role of an enduring educator. Such research findings show that parents' positive and active role cannot be underestimated. Second, the overexpectation of parents can push children towards frustration [121,122]. Parent mostly observes remarkable achievement on social media, so they also start demanding good grades from their children. With such pressure, students are easily frustrated, which negatively influences their academic outcome during cognitive activities, such as assignments, quizzes, and midand final-term examinations. So, the role of the parents should be supportive and motivational, which would help against unnecessary pressure.
To achieve a student performance prediction system, we need to consider parental involvement and the aforementioned other attributes, such as the cohabitation status of parents, the relationship among their parents, socioeconomic situation, and the number of children. Prediction systems need to quantize all these attributes to evaluate future student performance properly. If we look into literature studies, a minimal contribution can be evidenced toward mathematical modeling of student performance for a better educational system.

Emotional Factors.
Emotional attributes play a fundamental role in impacting student performance during cognitive tasks. e current study discusses severity levels of frustration, anxiety, depression, and stress. e impact of frustration is the natural part of learning as well as the engaging session (for references, see the literature review section). Such emotion is always found during comprehensive cognitive activities. Literature saturated with many qualitative findings focused on the statistical association between student performance and frustration. However, the study has not been evidenced a comprehensive approach to solve the challenges produced by frustration during cognitive tasks.
We need to analyze the performance of excellent, average, and at-risk students while mathematically modeling the relationship between institution-related attributes, students' emotional factors, and family-related attributes. Also, the teacher can help frustrated students' through collaborative exercises, group activities, and group assignments [123]. It will help students easily share their confusion and problems with group members to overcome their frustrations in a comprehensive learning environment. An individual can learn better in offline mode with face-to-face interaction as compared to online interaction [124]. Additionally, the COVID-19 outbreak has accelerated the influence of negative emotions on students' performance. COVID-19 has created a more critical situation for students' learning and adjusted them to the online environment with fewer resources. us, we are in dire need to evaluate the academic development of students while statistically associating the aforementioned factors and mathematically modeling the proposed relationship to prepare for the critical situation [125].

Gender Description.
In the literature review section, the study has shown that earlier studies statistically associated students' performance with emotional attributes and gender description. Students perform differently while considering aging and gender [126]. Both emotion and gender need to evaluate differently during cognitive activities. Literature studies are evidenced with many contributions on gender differences. ey show that different gender individuals perform differently during cognitive activities, solving assignments, attempting quizzes, and examinations. Earlier studies depict that gender difference is an independent biological factor whose magnitude is sometimes dependent on other factors such as cultures, socioeconomic condition, language, age, etc. Gender differences play a crucial role in influencing mental abilities and cognitive processing in mathematical tasks, physics, research, reading, and writing. ese issues create a big gap between male and female individuals, referred to as natural and biological differences.

Institution-Related
Factors. Different institutional factors are directly or indirectly involved in influencing students' performance.
ese factors include but are not limited to instructor teaching methodology, interaction with a student advisor, extracurricular activities in the institution, student complaint platform, the distance between the institution and students' residence, transport facility, and the behaviors of the friends. ese all factors have merits and demerits for student performance. e literature studies of psychology and data analysis have enormous contributions to student performance analysis; however, insignificant contributions have been reported in the form of algorithms and mathematical models in students' performance prediction.

Q3: What Is the Intensity of Research Findings in the field of Student Performance Prediction Systems Optimization?
Literature reported many challenges because the students' performance prediction overlaps psychology, data analysis, and mathematical and algorithmic contributions. e intensity of publications in the student performance prediction area is reported below.

Intensity of Psychological Findings.
As discussed earlier, we can find many psychological research contributions in the field of student performance analysis, which show that emotional attributes always affect students' performance during cognitive activities. So, to provide an efficient solution for student performance prediction, the study must need to evaluate the psychological findings that directly or indirectly focus on student performance evaluation.

Intensity of Data Analysis Findings.
Data analysis contribution provides a quantitative measurement for student performance prediction. Such research findings pave the way for an accurate mathematical model to better contribute to the performance prediction area of research.

Intensity of Students' Performance Prediction Systems.
e literature is also saturated with student performance prediction techniques focusing on students' performance prediction in critical cognitive tasks; nevertheless, these findings are not synchronized and linked toward a significant student performance model. So, the main objective of this review paper is to provide an effective platform for future researchers in student performance prediction. It will pave the way for an effective system to predict at-risk students and excellent student performances, which ultimately provides us with the opportunity to enhance their skills and performance.

Q4: Are the Findings of Psychological Studies, Data Mining, and Contribution in Algorithms Are Synchronized with Each Other for the Viability of Pilot Project?
e intensity of publications contributing to student performance prediction is quite good, but these contributions are not synchronized with each other to mathematically model emotional, family, and institution-related attributes. One of the main objectives of the current review is to highlight the lack of coordination and synchronization of the literature from different research fields. is review would allow future readers of deep learning to collaborate with other research fields.

Q5: How Synchronization and Coordination of Prior Psychological, Data Mining, and Algorithmic Findings Contribute to the Effective Educational System via Student Performance Prediction
Algorithm. Psychological literature produces both qualitative and quantitative findings in students' performance prediction; nevertheless, the data analysis field highlights the association among students' factors, i.e., emotional, family, and institutional attributes. If these findings are linked with the objective of qualitative data repositories and algorithms, then, we can move toward an efficient student performance prediction system. e psychological work produces accurate students' emotional data focusing on their performance. On the other hand, the data analysis field makes the meaningful statistical association and correlation information. e data analysis field of research provides a couple of tests to find the correlation between student emotional attributes and their performance, i.e., Pearson correlation and regression. ese tests verify the correlation among different factors.
We are in dire need to have the abovementioned psychological and data analysis findings to propose a comprehensive algorithm. Every part of the student performance prediction area of research is interlinked. e psychological result verifies the emotional change during the evaluations of the frustration, severity, anxiety, and stress. Second, the data analysis findings associate the student attributes. ird, the student performance prediction algorithm mathematically model the statistical association among the student influencing factors and their performance outcome.

Specific Keywords-Wise Publications.
is section intensively discusses the specific keyword-wise research output focusing on students' performance, emotional factors, and prediction algorithms. e list of keywords is illustrated in Table 1.
e study collected articles based on these keywords for further technical assessment. e specific domain for the technical evaluation includes but is not limited to new methods, modifications in prior work, data analysis, psychological findings, application analysis, review work, and comparison. e self-explanatory Table 1 illustrated the intensity of publications in the domain above using the list of keywords.

Yearly Publications.
Literature studies deliver thousand of research findings that directly or indirectly contribute to students' performance analysis and prediction. As illustrated in Figure 2, 37 published articles were evaluated (1990 to 1994). About 110 articles mainly focus on student performance and students' study-related factors assessment. ey have evaluated those factors that affect students' performance during cognitive activities (1995 to 1999).       Total number of  publications  Frustration  3  4  26  24  19  24  31  131  Frustration severity  0  0  13  33  4  35  14  99  Stress  2  2  8  19  13  19  16  79  Stress severity  0  0  18  18  9  11  21  77  Anxiety  19  12  14  17  9  14  6  91  Anxiety severity  0  2  10  22  21  14  24  93  Depression  0  0  12  20  23  12  21  88  Parents' influence  4  3  13  11  6  25  16  78  Distance from  home and school  3  3  18  13  13  13  19  82   Mobile game  2  2  27  25  17  12  24  109  Outdoor game  0  0  19  7  20  13  7  66  Indoor game  0  0  22  17  21  10  10  80  Watching TV  0  0  13  10  24  18  16  81  Students social  network  8  12  17  19  7  24  8  95   Gender  3  3  10  14  15  15  16  76  Parents  cohabitation status  0  0  11  6  15  18  14  64   Parent service  2  4  5  15  11  3  20  60  International  students  0  0  5  10  11  7  15 48  3  3  TSBCP  3  3  AUD  5  5  CPU  3  2  5  DDC  1  3  4  EDMD-APS  2  2 10 Computational Intelligence and Neuroscience   Computational Intelligence and Neuroscience 13 Computational Intelligence and Neuroscience Predicting student success using data generated in traditional educational environments DMA-SD Data mining application on students' data EDU-DMPC Educational data mining for prediction and classification of engineering students achievement PSP-CA A comparative analysis of techniques for predicting student performance PSS-CF Predicting students success in courses via collaborative filtering DMM-SC Data mining models for student careers PMSA Blending measures of programming and social behavior into predictive models of students achievement in early computing courses QACL Quantitative approach to collaborative learning HSE Will teachers receive higher student evaluations by giving higher grades and less course work? PPM-AS Student performance prediction model for early-identification of at-risk students in traditional classroom settings RAE Regression analysis by example ICASP Mining the impact of course assignments on student performance PSP-TDF Predicting student performance in an ITS using task-driven features SSC Soft subspace clustering of categorical data with probabilistic distance EDPLC Early detection prediction of learning outcomes in online short-courses via learning behaviors KPS-EDU Tracking knowledge proficiency of students with educational priors PSP-NBT Exploration of classification using NB tree for predicting students' performance SMA Student modeling approaches: a literature review for the last decade SMCS An ontological approach for semantic modeling of curriculum and syllabus in higher education PSP-LMS Predicting student performance from LMS data TLR Organizing knowledge syntheses: a taxonomy of literature reviews ARICM Analysis of academic results for informatics course improvement using association rule mining PSP-ALA Predicting student performance using advanced learning analytics SARL Seeding the survey and analysis of research literature with text mining EDM A systematic review of educational data mining TRI-PAP Do the timeliness, regularity, and intensity of online work habits predict academic performance? PSP-PA Predicting student performance using personalized analytics AWP Automated analysis of aspects of written argumentation TSBCP Predicting performance form test scores using back propagation and counter propagation AUD e text mining handbook: advanced approaches in analyzing unstructured data,cambridge CPU Cell phone usage and academic performance DDC Learning analytics: drives, developments and challenges EDMD-APS Educational data mining discovery standards of academic performance by students A model to predict low academic performance at a specific enrollment using data mining PSP Predicting students performance in educational data mining NSP-KDHED A new student performance analysing system using knowledge discovery in higher educational databases.

MLM
Comparison of machine learning methods for intelligent tutoring systems ID-CS Individual differences related to college students' course performance in calculus ‖ SAP Student academic performance prediction by using a decision tree algorithm. PP-PSP Performance prediction based on particle swarm optimization PSP-M Poverty and student performance in Malaysia PA-PS Physical activity is not related to performance at school PF e power of feedback, review of educational research IDF-SAP Identifying key factors of student academic performance by subgroup discovery SC-NF Student classification for academic performance prediction using neuro fuzzy in a conventional classroom OEP-TRF Online education performance predication via time-related features PCS Programming content semantics: an evaluation of visual analytics approach SVA-PC Semantic visual analytics for today's programming courses PSL A systematic review of studies on predicting student learning outcomes using analytics SAP-EDC Predicting student academic performance in an engineering dynamics course: a comparison of four types of predictive mathematical models PRD-AP Predicting student's academic performance: comparing artificial neural network, decision tree, and linear regression SSP-CL Analyzing student spatial deployment in a computer laboratory QE-ELC Quality enhancement for e-learning courses: the role of student feedback GRP-OEWB Improving accuracy of students' final grade prediction model using optimal equal width binning and synthetic minority over-sampling technique SP-DMC Student performance prediction by using data mining classification algorithms PPRD-DT Performance prediction of engineering students using decision trees SUR-MSR A survey and taxonomy of approaches for mining software repositories in the context of software evolution SPRD-ARMBA A review and performance prediction of students' using an association rule mining based approach EXP-HPF Exploring the high potential factors that affects students' academic performance IPT-SP Analysing the impact of poor teaching on student performance DM-ETSP Data mining based analysis to explore the effect of teaching on student performance SPP-DL Gritnet: student performance prediction with deep learning DM-E Data mining and education PSM-HOU Predicting students marks in hellenic open university PSP-ML Predicting postgraduate students' performance using machine learning techniques PA-EDM Review on prediction algorithms in educational data mining LS-PRDE Literature survey on student's performance prediction in education using data mining techniques PRD-AP Predicting student academic performance HSC-SA Online self-paced high-school class size and student achievement PRI-MPP Predictor relative importance and matching regression parameters SE-OES Finding similar exercises in online education systems FCD-EP Fuzzy cognitive diagnosis for modeling examine performance EB-PSP An ensemble-based semi-supervised approach for predicting students' performance MSM-ENB Measuring the (dis-) similarity between expert and novice behaviors as serious games analytics 16 Computational Intelligence and Neuroscience Abbreviation Acronym M-KME Mining for topics to suggest knowledge model extensions EPRD-BL Applying learning analytics for the early prediction of students' academic performance in blended learning MA-TE Whose feedback? A multilevel analysis of student completion of end-of-term teaching evaluations PRD-SF-GP Predicting student failure at school using genetic programming and different data mining approaches with high dimensional and imbalanced data SE-UT Students' evaluations of university teaching: Dimensionality, reliability, validity, potential biases and usefulness SE-DRVPU Students' evaluations of university teaching: Dimensionality, reliability, validity, potential biases and usefulness PRD-DFA Predicting student outcomes using discriminant function analysis AANL-PISA An overview of using academic analytics to predict and improve students' achievement: a proposed proactive intelligent intervention

INT-INF
Constructing interpretive inferences about literary text: the role of domain-specific knowledge PRD-GR Predicting grades ESEG-AP Early segmentation of students according to their academic performance: a predictive modeling approach SAG-EDM A framework for smart academic guidance using educational data mining MSD-PRD Mining students' data for prediction performance PRI-SRMA Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement SRS-AL A semantic recommender system for adaptive learning SET-IFP Students evaluating teachers: exploring the importance of faculty reaction to feedback on teaching SRL-HYPM Self-regulated learning with hypermedia: the role of prior domain knowledge SAP-DM Modeling and predicting students' academic performance using data mining techniques LAS-TEL Lexical analysis of syllabi in the area of technology enhanced learning DMKMS Student data mining solution-knowledge management system DSS-LE Decoding student satisfaction: how to manage and improve laboratory experience FGCAC Student ability best predicts final grade in a college algebra course SAPM Student academic performance monitoring and evaluation DM-PSP Data mining approach for predicting student performance OPCA Optimizing partial credit algorithms HESSP-PP Is alcohol affecting higher education students' performance: searching and predicting pattern IOMC Towards the integration of multiple classifier pertaining to the student's performance prediction DM-CRTL A data mining view on classroom teaching language DM-ED Application of data mining in educational databases for predicting academic trends and patterns FGSK-SP Using fine-grained skill models to fit student performance EDM-ARW Educational data mining: a survey and a data mining-based analysis of recent works GP-SSM Grade prediction with course and student specific models FENTP Feature extraction for next-term prediction of poor student performance TE-LMSF Teaching evaluation using data mining on moodle LMS forum RGTE e role of gender in students' ratings of teaching quality in computer science and environmental engineering DOF-DTT Drop out feature of student data for academic performance using decision tree techniques P-CSI Programming: predicting student success early in CSI DTDM Decision trees and decision-making PSP-SDMA Predicting student performance: a statistical and data mining approach SAS A sentiment analysis system to improve teaching and learning ODF-AFQP Ontology driven framework for assessing the syllabus fairness of a question paper PSP-OLDF Predicting students' final performance from participation in on-line discussion forums EDM-S Educational data mining: a survey from 1995 to 2005 EDM-RSA Educational data mining: a review of the state of the art ASP-DCBC Analyzing student performance using sparse data of core bachelor courses CSP-LCV Centralized student performance prediction in large courses based on low-cost variables in an institutional context EEDM-IPC Evaluating the effectiveness of educational data mining techniques for early prediction of students' academic failure in introductory programming courses PSP-C Prediction of students' academic performance using clustering PSP-DMT A review on predicting students' performance using data mining techniques WUGC Web-based undergraduate chemistry problem-solving: the interplay of task performance, domain knowledge and websearching strategies SEDM-PSP A survey on various aspects of education data mining in predicting student performance LAEDM-CC Learning analytics and educational data mining: towards communication and collaboration PSP-EDT Predictive modeling of students performance through the enhanced decision tree TQSA-ES What is the relationship between teacher quality and student achievement? An expletory study PMTP A predictive model for standardized test performance in Michigan schools DFUS Determination of factors influencing the achievement of the first-year university students Abbreviation Acronym SPP-CS Next-terms student performance prediction: a case study MED-CS Mining educational data to improve students' performance: a case study IAPP Improving academic performance prediction by dealing with class imbalance MM-SN Proposing stochastic probability-based math model and algorithms utilizing social networking and academic data TQ-CS Teaching quality matters in higher education: a case study MA-FTE Meta-analysis of faculty's teaching effectiveness: student evaluation of teaching ratings and student learning FGAM Analysis of the impact of action order on future performance: the fine-grain action model MRF-CA Map-reduce framework based cluster architecture for academic students' performance prediction GSM Google Scholar coverage of a multidisciplinary field OCM e opportunity count model: a flexible approach to modeling student performance LRMP Predicting students' performance in final examination using linear regression and multilayer perceptron IDK Fast searching for information on the internet to use in a learning context: the impact of domain knowledge EDM Educational data mining acceptance among undergraduate students ILA-EDM Participation-based student final performance prediction model through interpretable genetic programming: integrating learning analytics, educational data mining and theory RPP Improving retention performance prediction with prerequisite skill features SP-RBFNN&PCA Predicting honors student performance using RBFNN and PCA method SP-MLR&PCA Predicting students' academic performance using multiple linear regression and principal component analysis

NCAS-COVID19
Negative emotions, cognitive load, acceptance, and self-perceived learning outcome in emergency remote education during COVID-19 Imp-COVID19 e impact of COVID-19 on education insights from education at a glance 2020 SS-PPP-DM Study on student performance estimation, student progress analysis, and student potential prediction based on data mining A-EDM-TD Application of educational data mining approach for student academic performance prediction using progressive temporal data RPSP-DMT A review on predicting students' performance using data mining techniques SPP-CL Student performance analysis and prediction in classroom learning: a review of educational data mining studies ER-KCP Exercise recommendation based on knowledge concept prediction SDP Student dropout prediction EDP-DM Early dropout prediction using data mining: a case study with high school students PAP-SH Predicting academic performance by considering student heterogeneity HMRS Helping university students to choose elective courses by using a hybrid multicriteria recommendation system with genetic optimization IGR-PSP Inductive Gaussian representation of user-specific information for personalized stress-level prediction PSPP-ML Pre-course student performance prediction with multi-instance multi-label learning ECE-RL What students want? Experiences, challenges, and engagement during emergency remote learning amidst COVID-19 crisis 18 Computational Intelligence and Neuroscience ATI-F "Affect-targeted interviews for understanding student frustration", in international conference on artificial intelligence in education CCI-OC Common challenges for instructors in large online course: strategies to mitigate student and instructor frustration ETES-COVID19 Effective teaching and examination strategies for undergraduate learning during COVID-19 school restrictions TFL-SF Teacher feedback literacy and its interplay with student feedback literacy OC-BL Challenges in the online component of blended learning: a systematic review NP-PSP Feature extraction for next-term prediction of poor student performance SPP-BL Student performance prediction based on blended learning RSNL Robust student network learning DN-CS Deep network for the iterative estimations of students' cognitive skills PR-MS Parents' role in the academic motivation of students with gifts and talents DSF-HB Detecting student frustration based on handwriting behavior VFP-C e validity of a frustration paradigm to assess the effect of frustration on cognitive control in school-age children EAK-P Ekt: exercise-aware knowledge tracing for student performance prediction SP-EG-MM Predicting student performance in an educational game using a hidden Markov model LMS-CAP Massive lms log data analysis for the early prediction of course-agnostic student performance FDG Frustration drives me to grow BFE Between frustration and education: transitioning students' stress and coping through the lens of semiotic cultural psychology AD-CS Automatic discovery of cognitive skills to improve the prediction of student learning SP-ALA Predicting student performance using advanced learning analytics TVL-CA Time-varying learning and content analytics via sparse factor analysis EAG-CSC Emotions, age, and gender based cognitive skills calculations ML-CSC Machine learning based cognitive skills calculations for different emotional conditions SP-DM-LAT Predicting student performance using data mining and learning analytics techniques: a systematic literature review S-GC Should I grade or should I comment: links among feedback, emotions, and performance MR-PCQ Modeling the relationship between students' prior knowledge, causal reasoning processes, and quality of causal maps MPA-M A multilayer prediction approach for the student cognitive skills measurement MCA-E A meta-cognitive architecture for planning in uncertain environments T-PR e influence of teacher and peer relationships on students NS-SE National Society for the Study of Education ARFE Automatically recognizing facial expression: predicting engagement and frustration CSMA A biologically inspired cognitive skills measurement approach NT-PPCS A novel technique for the evaluation of posterior probabilities of student cognitive skills MSG-IC Medical student gender and issues of confidence GD-ATC Gender differences in student attitudes toward computers GD-AT-SCI Gender differences in student attitudes toward science: a meta-analysis of the literature from 1970 to 1991 LS-ESP-R A longitudinal study of engineering student performance and retention III. Gender differences in student performance and attitudes GD-SE Gender differences in student ethics: Are females really more ethical? Gender differences in teacher-student interactions in science classrooms GD-AT-IT Gender differences in attitudes towards information technology among Malaysian student teachers: a case study at University Putra Malaysia GD-RC Gender differences in the response to competition GD-LTS Gender differences in the learning and teaching of surgery: a literature review SG-TM-CAP Student gender and teaching methods as sources of variability in children's computational arithmetic performance GDSL Gender difference and student learning GD-MS-SL Gender difference in student motivation and self-regulation in science learning: a multigroup structural equation modeling analysis GD-MR Gender differences in the influence of faculty-student mentoring relationships on satisfaction with college among African-Americans DSS-CP Differences of students' satisfaction with college professors: the impact of student gender on satisfaction GD-TET Gender differences in teachers' perceptions of students' temperament, educational competence, and teachability GD-HSS Gender differences in factors affecting academic performance of high school students TP-MA Influence of elementary student gender on teachers' perceptions of mathematics achievement Abbreviation Acronym GD-NCS Gender differences in alcohol-related non-consensual sex, cross-sectional analysis of a student population GD-SP-EC Gender differences in students' and parents' evaluative criteria when selecting a college SSG Social influences, school motivation, and gender differences: an application of the expectancy-value theory GD-DSS Gender differences in the dimensionality of social support ETP-SSA Early teacher perceptions and later student academic achievement GES-E Gender, ethnicity, and social cognitive factors predicting the academic achievement of students in engineering PSD Predicting students drop out: a case study IQ-PAP Self-discipline outdoes IQ in predicting academic performance of adolescents TSI-SSC Observations of effective teacher-student interactions in secondary school classrooms: predicting student achievement with the classrooms assessment scoring system-secondary BFP-MA Role of the big five personality traits in predicting college students' academic motivation and achievement ESF-SS Using emotional and social factors to predict student success FPP-AUS Who succeeds at university? Factors predicting academic performance in first-year Australian university students ACA Predicting academic achievement with cognitive ability AAGT Advancing achievement goal theory: using goal structures and goal orientations to predict students' motivation, cognition, and achievement SLC-A Short-term and long-term consequences of achievement goals: predicting interest and performance over time RHAS Role of hope in academic and sports achievement PSO-LPS Prediction of school outcomes based on early language production and socioeconomic factors Prediction of at-risk students for special treatment and counseling sessions.
Mathematically model emotional attributes, family issues, study schedules, and institutional attributes all together to develop a significant prediction system. If students cannot achieve an excellent academic score, then the performance prediction system assists students in observing the main reason behind the low performance.
If the prediction system considers a large number of influential factors, then the academic achievement of excellent can also be advanced.
Advance students' academic achievements.
Modulates the relationship between behavior and students' performance Monitor students' behavior such as interaction and attitude towards teacher, seriousness, and unseriousness in the classroom 2 What are the factors that can optimize student performance prediction? ey include but are not limited to familyrelated factors, emotional factors, gender description, and institution-related factors.
Initiate pilot projects with an assumptionbased dataset. e assumptions should be based on earlier studies of psychology, data analysis, and data mining.
Emotional factors, such as frustration, anxiety, stress, and depression.
Analyze the performance of at-risk students while mathematically modeling the association among students' emotional, family, and institution-related attributes.
Quantize family factors, i.e., parents' positive and negative roles, including overexpectation of parents and positive involvement of parents in children's daily cognitive activities.
Perform factorization of gender because earlier studies depict that gender difference magnitude is sometimes dependent on other factors such as cultures, socioeconomic condition, language, age, etc. Literature studies are evidenced with many contributions to gender differences. ey show that different gender individuals perform differently during cognitive activities, solving assignments, attempting quizzes, and examinations studies.
Explore instructor teaching methodology, interaction with a student advisor, extra curriculum activities in the institution, student complaint platform, the distance between the institution and students' residence, transport facility, and the behavior of the friends. Different institutional factors directly or indirectly influence students' performance.
publish the most highly cited research papers. It involves overall 1497 publications, which are selected after searching on Google Scholar. ese studies were evaluated upon their relevant findings, such as new methods, modified approaches, statistical findings, and psychological results (for more information see Table 2 and Figure 3). e major part of this review was to assess the existing students' performance prediction approaches. So, the study analytically assessed new students' performance prediction measures, modifications in state-of-the-art techniques, and comparative analysis. Additionally, Figure 4 and Table 3 represent the detailed domain-wise and factors-wise analysis.

Potential Future Challenges
A large number of factors are involved in influencing students' performance; therefore, the prediction system needs to be optimized to consider the impacts of different human factors categories. Such factors categories include but are not limited to emotional attributes, study schedule, family attributes, and institutional attributes. Each category consists of multiple factors impacting students' performance, either negatively or positively. In the literature section, the study provides a detailed discussion of these factors.

Potential Pilot Projects Based on the Assumption-Based
Dataset. e comprehensive synchronization between the earlier studies is still a black box, which increases systems' dependency on a real-world dataset. e importance of a realworld dataset cannot be avoided; however, the data collection process is time-consuming and need a list of human resources. It delays the optimization of existing approaches, such as modeling students' emotional attributes. e data collection process could have various anomalies if the researcher does not follow the analysis of earlier studies. e earlier studies offer excellent opportunities to understand the effectiveness of emotional attributes for optimization. erefore, pilot projects perform key roles in optimizing the existing students' performance prediction systems. ey provide useful ideas during the data collection process. ey also pave the way for an assumption-based dataset to prove the viability of novel ideas in students' performance prediction.

Additional Points of Earlier Studies
Data analysis findings explore the hidden patterns and statistical correlation between students' performance and influential factors. Such opportunities introduce new challenges for students' performance prediction systems, e.g., conditional probabilities, correlation, and inferencing. Also, data mining studies are evidenced with many findings in students' performance prediction area of research; nevertheless, they have different limitations, e.g., lack of in-depth investigation of students' performance based on selected study-related factors, limited scalabilities, limited dataset, and inadequate qualitative approach of data analysis and psychological studies. What is the intensity of research findings in the field of student performance prediction systems optimization?
Intensity of psychological findings ese findings are not synchronized and linked toward a significant student performance prediction model.

Intensity of data analysis findings
So, the main challenge is to provide an effective platform where future researchers can collaborate and synchronize the prior findings. Also, pilot projects based on the assumption-based dataset are highly recommended. Successful pilot project implementation will pave the way for quick optimization of existing systems. Intensity of students' performance prediction systems 4 Are the findings of psychological studies, data mining, and contribution in algorithms synchronized with each other for the viability of the pilot project?
e intensity of publications contributing to student performance prediction is quite good, but these contributions are not synchronized with each other.
Mathematically model emotional, family, and institution-related attributes.

5
How do synchronization and coordination of prior psychological, data mining, and algorithmic findings contribute to the effective educational system via student performance prediction algorithm?
Every part of the student performance prediction area of research is interlinked. e psychological result verifies the emotional change during the evaluations of the frustration, severity, anxiety, and stress. e data analysis findings associate the student attributes. e student performance prediction algorithm mathematical model the statistical association among the student influencing factors and their performance outcome.
If these findings are linked with the objective of qualitative data repositories and algorithms, then, we can move toward an efficient student performance prediction system.

Computational Intelligence and Neuroscience 21
Finally, the review shows that various prior students' performance prediction methods have been proposed in the last decade; however, meagre studies have highlighted the basic need for synchronization among the abovementioned field's contributions. erefore, this review provides an exclusive picture of the future challenges in students' performance prediction (see Table 1 to 4 and Figure 2 to 4). On one hand, Table 4 depicts the intensity of various optimization techniques, review works, and new students' performance prediction methods. On the other hand, Table 5 represents the acronyms of the selected studies. Remarks and recommendations against each research question are given in the self-explanatory Table 6.

Conclusions
e proposed review highlights the potential research opportunities to optimize the students' performance prediction systems while exploring earlier contributions of different research fields, i.e., cognitive computing, data mining, data analysis, and psychology. e previous studies are still limited in synchronization between the existing contributions of various fields, which negatively impacted the mathematical modeling of emotional attributes. It increased the systems' dependencies on real-world datasets. us, to investigate the potential challenges thoroughly, the study is split into three sections.
1. e data mining discoveries, psychological findings, and data analysis results are examined. 2. e study performs a domain-wise investigation of the existing methods focusing on students' performance prediction, i.e., the domain includes new students' performance prediction techniques, modifications in existing techniques, and comparisons analysis. 3. Eventually, future direction and potential pilot project viability are highlighted.
Data Availability e screening data are available from the corresponding author, upon reasonable request.

Conflicts of Interest
e authors declare that they have no conflicts of interest.