TOURISM ATTITUDE CLASSIFICATION SYSTEM USING NAIVE BAYES TECHNIQUE

Abstract

The purposes of this research study are to develop the program for distinguishing the positive and negative tourism attitudes by using the Naive Bayes technique and to evaluate the efficiency of the tourism attitude system by using accuracy matrices, recognitions, F-measure and correctness. The data were Thai tourists' reviews about domestic tourism from Google Map and TripAdvisor website. It was found that the program could analyze and distinguish the feelings from general texts. However, there might be problems about sentences with unclear sentences and phrases such as not really bad. These texts were problematic for cutting words and interpreting by the program. Consequently, scores are incorrect. By evaluating the efficiency of the system with 120 different documents for creating models; the recognition, accuracy and F-measure values are 80, 100 and 120, respectively. The averages from the three models are as follows. The average recognition is 0.86. The average accuracy is 0.80. The average F-measure is 0.80. The average correctness is 0.80.