Perbandingan Algoritma Naïve Bayes Dan K-Nearest Neighbor pada Analisis Sentimen Pembukaan Pariwisata Di Masa Pandemi Covid 19

Abstract

Covid-19 is a trending topic of conversation on social media today. one of the trending topics on social media in 2021 in the midst of the covid 19 pandemic related to tourism destinations that will reopen amid the increase in daily cases of covid 19 in a number of areas of Indonesia in particular. The implementation of the government's policy of opening tourist destinations has drawn various arguments or opinions among the public. especially social media users by providing arguments or tweets related to opinions through social media such as Instagram, Twitter, and Facebook. In this study, an analysis of the sentiments of the Indonesian people, especially social media users, was carried out regarding government policies related to the opening of tourism in the midst of the COVID-19 pandemic using the Naive Bayes and K-Nearest Neighbor (KNN) algorithms and knowing the comparison of the performance of the two algorithms. This dataset used is taken through various social media platforms with a data crawling process. The dataset used is a sql file. the number of datasets used to process data processing in the public sentiment analysis application about the opinion of opening tourism in the midst of the covid 19 pandemic on twitter as many as 570 tweets using 12 hashtags namely #Desabisa, #desawisata, #DiIndonesiaAja, #EraNewNormal and #LabuanBajo #MulaiDariDesa #PPKMDiPerpanjang #SobatParekraf #PlacesTourism #Bali Tours #Banyuwangi Tours #Jogja Tours. In the research conducted, the implementation of the results of the application of the Naive Bayes algorithm and the K-Nearest Neighbor (KNN) algorithm in the sentiment analysis of the opening of tourism destinations during the covid-19 pandemic was successfully carried out by producing the highest Naive Bayes accuracy rate of 75.53%, precision the highest positive is 71%, the highest negative precision is 25%, the highest positive recall is 99% and the highest negative recall is 14%. As for the K-Nearest Neighbor (KNN) algorithm, the highest accuracy rate is 48.66%, the highest positive precision is 69%, the highest negative precision is 14%, the highest positive recall is 69% and the highest negative recall is 28%. The level of accuracy in the comparison of the two algorithms shows that the Naive Bayes algorithm has the best performance with an accuracy level compared to the K-Nearest Neighbor algorithm in analyzing the sentiments of social media users about the opening of tourism during the COVID-19 pandemic.