La prédiction en apprentissage automatique, comment marche les algorithmes de classification des données : le cas de renforcement de capacité en TIC des étudiants de l’ENSup

In this paper, we describe the different steps to solve a prediction problem using machine learning algorithms, particularly classification algorithms. Our case study is the capacity building in ICT (Information and Communication Technology) of students at ENSup (Ecole Normale Supérieure). In our case study, we have a binary classification because we determine whether students need capacity building (yes) or not (no). A prediction in classification is generally done in five phases: the investigation phase, preprocessing, dataset separation, modeling and evaluation. The survey phase is used for the collection of raw data (primary data). In this article, we use a questionnaire of 101 questions for 250 students. The preprocessing phase is used to transform the raw data into a dataset. The dataset is divided into training and testing data in the separation phase. We use cross-validation separation. Only the training data will be used for creating the algorithm model in the modeling phase. The test data is used to evaluate the different algorithms in the evaluation phase. We use the following classification algorithms: Logistic Regression (LR), Naïve Bayes (NB.), Support Vector Machine (SVM), K-Nearest Neighbors (KNN) and Multilayer Perceptron (MLP). The results show that MLP and LR perform better than the others. Keywords: ICT, ENSup, Classification algorithms, Machine Learning, Data prediction