Publication

Abstract : The Bag of Words (BOW) technique is a Natural Language Processing technique that is applied to learn from the text. The basic idea of the Bag of Words is to take any text and count the frequency of words and place those unique words in a dictionary. This research is carried out to predict the causes of suicide in Nigeria by using the Bag of Words (BOW) technique together with three machine learning algorithms. The logistic regression algorithm, Random Forest classifier, and XGBoost classifier are the machine learning algorithms considered for this dataset. For feature extraction, the TfidfVectorizer has been used. The result gotten shows that the logistic regression classifier is suitable for the dataset used in this study, therefore, for similar datasets, the logistic regression together with TfidfVectorizer can be used. However, the logistic regression gave an accuracy score of 89% leaving an 11% chance of misclassification.