Dimensionality Reduction for Classification of Filipino Text Documents based on Improved BayesianVectorization Technique

Abstract

Dimensionality reduction of feature vector size plays a vital role in enhancing the text processing capabilities to reduce the size of the feature vector used in the mining tasks to achieve a higher classification accuracy. While dimensionality reduction for text classification is becoming a great area of research in most languages, Filipino documents have received little or no attention from researchers. Thus, this paper addresses the issue of dimensionality reduction in representing relevant data from Filipino texts using an improved Bayesian vectorization technique. To validate the effectiveness of improved Bayesian vectorization, the model was compared to the Term Frequency and Inverse Document Frequency (TF-IDF) method. The outcomes are presented using standard measures such as precision, recall, f-score, and accuracy. The results revealed that the improved Bayesian vectorization has significantly better results having 98% classification accuracy compared to 76% classification accuracy of the TF-IDF vectorization technique.

A Data Analytic Approach in the Thematic Classification of the Reasons and Perspectives of Adolescents’ Social Media Engagement

“…Documents in the bag of words containing no words were removed as well as their entry labels. A support vector machines (SVM)-based supervised classification model was used using the word frequency counts from the bag-of-words model and the labels [21], [22]. A multiclass linear classifier specifies the counts of the bag-of-words model to be the predictor, and the event type labels to be the response.…”