APPLYING Machine Learning in Predicting Probable Senior High School Strand Basis for Career PLANNING

Abstract

Three popular algorithms were considered to create the Strand Prediction System: the Naïve Bayes, Deep Neural Network, and Random Forests. The three algorithms in terms of accuracy were DNN (84.70%), Random Forest (95.25%), and Naïve Bayes (52.9%). Random Forest's ensemble technique ensured consistent and reliable performance while reducing overfitting. To enhance the dataset for the model creation, the SMOTE (Synthetic Minority Over-sampling Technique) was used to balance the distribution of the strand instances, thereby creating an acceptable dataset, which was used in the WEKA application to create the models. By doing so, the accuracy rate was increased. The system prototype passed all the tests, and the high average score of 4.56 shows that users think it is valuable and easy to use. To top it all off, it was well-received. However, issues with the results' demonstrability revealed areas that may be improved. The researcher secured the authorities' approval before collecting data from different schools and administering the adopted TAM2 survey questionnaire.