Dita Sekar Asri, - (2024) ANALISIS SENTIMEN MASYARAKAT DI MEDIA SOSIAL TWITTER MENGENAI LAYANAN JASA SICEPAT MENGGUNAKAN METODE NAÏVE BAYES CLASSIFIER. Skripsi thesis, Universitas Islam Negeri Sultan Syarif Kasim Riau.

Preview	Text DITA SEKAR ASRI 11950125032 TA2 Tanpa BAB 4.pdf Download (2MB) | Preview
	Text (BAB 4) DITA SEKAR ASRI 11950125032 TA2 BAB 4.pdf - Published Version Restricted to Repository staff only Download (1MB)

Abstract

Currently, many expedition services have been established because delivery services are really needed as a facility for sending goods to buyers' addresses due to the large number of online shopping transactions. SiCepat provides service and comfort to customers, but the service provided by SiCepat does not necessarily provide satisfaction to all customers. Customers can provide positive or negative comments on the services that SiCepat provides. Indonesian customers or people in expressing their satisfaction often involve public services that are easily accessed by many people, namely social media, one of which is Twitter. These tweets or comments can be collected as a data source in text mining research and processed into information. In this research, the application of text mining techniques used for tweet text classification is a machine learning method with the Naïve Bayes Classifier algorithm. This research uses data crawling techniques by collecting a dataset of 3000 tweets with positive and negative sentiment classes. Based on the results of the tests carried out, the highest accuracy was obtained with a test ratio of 90% train and 10% test with a fairly good accuracy value of 90.66%, a recall value of 11.11% and a precision value of 14.28%, then added with the K-fold Cross Validation test which gave the highest accuracy results in the third trial, namely a value of 95.98%, a recall value of 21.42% and a precision value of 75.0% and an average accuracy of 92,201%,. It can be concluded that the Naive Bayes Classifier has succeeded in classifying public sentiment on social media towards SiCepat services. Keywords: SiCepat, Classification, k-fold Cross Validation, Naïve Bayes Classifier, Text Mining, Twitter

Actions (login required)

View Item