ESTIMATING CUSTOMER SATISFACTION USING NAÏVE BAYES MACHINE LEARNING MODEL

Joseph B. Campit
Page No. 1-6

The study investigated the classification performance of Naive Bayes (NB) Machine Learning Model in estimating customer satisfaction. Likewise, it also determined the effect of applying different model configurations such as n-gram, stop words, and stemming on the classification performance of the model. Sentiment analysis was employed to analyze useful information from the unstructured responses of the respondents. The dataset consists of Tagalog and English words that were manually annotated and were randomly selected and assigned to the training and testing dataset. The general framework of the study consists of data preprocessing, modelling, and model comparison. Finding revealed that the highest classification performance of NB is attained using NB trigram with stop words removal.
Keywords : Machine Learning, Naïve Bayes, Sentiment Analysis, Text Analytics, classification performance