A Study of Feature Selection Algorithms for Predicting Students Academic Performance

Zaffar, Maryam; Hashmani, Manzoor Ahmed; Savita, K.S.; Rizvi, Syed Sajjad Hussain

doi:10.14569/ijacsa.2018.090569

Cited by 55 publications

(37 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Update the position of the current search agent by equation (5) end for Update A, a, and c Find the fitness function of each search agent Update x α , x β , and x δ set t=t+1 end while return x α…”

Section: Of Iterations) For Each Search Agentmentioning

confidence: 99%

Untitled

2020

Ann Proteom Bioinform

View full text Add to dashboard Cite

In the classifi cation of cancer data sets, we note that they contain a number of additional features that infl uence the classifi cation accuracy. There are many evolutionary algorithms that are used to defi ne the feature and reduce dimensional patterns such as the gray wolf algorithm (GWO) after converting it from a continuous space to a discrete space. In this paper, a method of feature selection was proposed through two consecutive stages in the fi rst stage, the fuzzy mutual information (FMI) technique is used to determine the most important feature selection of diseases dataset through a fuzzy model that was built based on the data size. In the second stage, the binary gray wolf optimization (BGWO) algorithm is used to determine a specifi c number of features affecting the process of classifi cation, which came from the fi rst stage. The proposed algorithm, FMI_BGWO, describes effi ciency and effectiveness by obtaining a higher classifi cation accuracy and a small number of selected genes compared to other competitor algorithms.How to cite this article: Noori NM, Qasim OS. Improving cancer diseases classifi cation using a hybrid fi lter and wrapper feature subset selection. Ann Proteom Bioinform. 2020; 4: 006-0011.

show abstract

Section: Of Iterations) For Each Search Agentmentioning

confidence: 99%

Untitled

2020

Ann Proteom Bioinform

View full text Add to dashboard Cite

show abstract

“…Zaffar et al [12] exhibited an examination of the presentation of filter feature selection algorithms and classification algorithms on two different student datasets. The outcomes we get from different FS algorithms and classifiers on two student datasets with different number of features will assist researchers to catch the best combinations of filter feature selection algorithms and classifiers.…”

Section: Literature Reviewmentioning

confidence: 99%

Ensemble Swarm based Feature Selection (ESFS) and Ensemble Three Classifiers (ETCS) to Predict Student’s Academic Performance

2019

ijeat

View full text Add to dashboard Cite

Recently Educational Data Mining (EDM) has attracted many researchers in recent years. Many techniques of data mining are formulated to generate the techniques of the knowledge that is hidden within the educational data. The knowledge which is extracted aid the educational institutions to enhance the teaching process and learning methods. These improvements enhance the student performance and the performance of overall outputs. In EDM, Feature Selection (FS) plays a significant role in the improvement of quality of the models used for the purpose of prediction of educational datasets. Single feature selection algorithms do not render enhanced results of prediction. In this proposed work, Ensemble Swarm based Feature Selection (ESFS) and Ensemble Three Classifiers (ETCs) is formulated to classify the performance of students based on the selected features. This work concentrates on ESFS techniques are formulated to select the important and intrinsic features before the process of classification, ETCs are proposed. The samples are selected from the knowledge repository, which is initially pre-processed by means of Min Max Normalization (MMN) and Z Score Normalization (ZCN) method. Then the selected attributes from the technique called Ensemble Swarm based Feature Selection (ESFS) are combined to the learner’s communication together with e-learning management system. ESFS algorithm fuses the Fuzzy Membership Genetic Algorithm (FMGA) and Improved Clonal Selection Algorithms (ICSAs). Also, Ensemble Three Classifiers (ETCs) is identified for the prediction of students’ performance by combining the qualifiers like Adaptive Neuro Fuzzy Inference System (ANFIS), Support Vector Machine (SVM) classifier and Decision Tree (DT). A widespread ensemble approach namely Bagging is utilized to combine all the results of three classifiers. The results that are obtained are found to have strong relationship among the learner’s behaviors and their academic achievement.

show abstract

“…Number of enrolled students in Higher education institutions has increased rapidly equivalent to 176% increase in the year 2017-18 with female 76% and males 26%[Times of Oman 2017] . Female students are coming forward for higher education in Sultante of Oman in the recent years than male students [7]. As per the statistics in 2014-15 57.7% female students have enrolled in higher education' [8].…”

Section: Introductionmentioning

confidence: 98%

Academic Performance Based on Gender using Filter Ranker Algorithms - An Experimental Analysis in Sultanate of Oman

Krishnan¹,

Beegum²,

P.C³

2019

IJITEE

View full text Add to dashboard Cite

Education molds the future society. Student profile analysis in higher education system in sultanate of Oman reveals that drop out cases of students are tremendously increasing for the last few years. Learning environment plays a vital role in providing appropriate teaching methodologies to motivate the skills of the students. Variations in academic performance can be observed based on different educational indicators such as gender, social, economical, cultural and community characteristics. This paper tries to conduct an analysis on the existing data based on educational data mining and tries to make a classification based on gender which helps to adapt necessary teaching methodology to improve the student performance. A data set of 400 students from three colleges in Sultanate of Oman in three consecutive years is taken as case study for this analysis. Since student's performance is classified as per the data model based on their gender, equal number of male and female data are considered. Data consists of 12 attributes with result as output. Classification of students performance has been achieved using the classification algorithms in WEKA tool. Irrelevant features are eliminated to select subset of input variables in feature selection (FS) algorithm. This is active to improve learning efficiency, predictive accuracy and reduce complexity of learned results. The paper demonstrates gender as the priority attributeto create best data model.

show abstract

A Study of Feature Selection Algorithms for Predicting Students Academic Performance

Cited by 55 publications

References 0 publications

Untitled

Untitled

Ensemble Swarm based Feature Selection (ESFS) and Ensemble Three Classifiers (ETCS) to Predict Student’s Academic Performance

Academic Performance Based on Gender using Filter Ranker Algorithms - An Experimental Analysis in Sultanate of Oman

Contact Info

Product

Resources

About