Analisis Sentimendan Pemodelan Topik Pariwisata LombokMenggunakan Algoritma Naive Bayesdan Latent Dirichlet Allocation

Analisis Sentimendan Pemodelan Topik Pariwisata LombokMenggunakan Algoritma Naive Bayesdan Latent Dirichlet Allocation

sentiment analysis, probabilistic computing, machine learning, tourism

ombok Island is one of the favorite tourist destinations. Various topics and comments about Lombok tourism experience through social media accounts are difficult to manually identify public sentiments and topics. The opinion expressed by tourists through social media is interesting for further research. This study aims to classify tourists' opinions into two classes, positive and negative, and topics modelling by using the Naive Bayes method and modeling the topic by using Latent Dirichlet Allocation (LDA). The stages of this research include data collection, data cleaning, data transformation, data classification. The results performance testing of the classification model using Naive Bayes method is shown with an accuracy value of 92%, precisionof 100%, recall of 84% and specificity of 100%. The results of modeling topics using LDA in each positive and negative class from the coherence value shows the highest value for the positive class was obtained on the 8th topic with a value of 0.613 and for the negative class on the 12th topic with a value of 0.528. The use of the Naive Bayes and LDA algorithms is considered effective for analyzing the sentiment and topic modelling for Lombok tourism

Ni Luh Putu Merawati1, Ahmad Zuli Amrullah2, Ismarmiaty3

https://jurnal.iaii.or.id/index.php/RESTI/issue/view/20

Universitas Bumigora

20 Februari 2021

Fajar bagus W

PDF

Indonesia

Text