Text

ANALISIS HARGA SAHAM PT ASTRA INTERNASIONAL Tbk MENGGUNAKAN DATA BURSA EFEK INDONESIA DALAM JANGKA WAKTU PENDEK MENGGUNAKAN METODE NAЇVE BAYES DAN DECISION TREE-J48

Share :
• 
• 
• 
• 
• 
• 
• 

ABSTRACT ANALYSIS Of The SHARE PRICES Of PT ASTRA INTERNATIONAL Tbk, USING DATA FROM The INDONESIA STOCK EXCHANGE In A SHORT PERIOD Of TIME USING The METHOD Of NAЇVE BAYES And DECISION TREE-J48 Heima Vera Yanti Sitorus NIM: 1481006 Stock price prediction is a very necessary action as a first step to taking the decision to determine when to transact a sale and purchase of existing shares in Indonesia stock exchange. Stock price prediction is done using two methods are compared, Naїve Bayes method and the method of Decision Tree-J48. The amount of data used is as much as 1,195 and data used as the data for testing is as much as 20% or 239 the data. As a result of the level of accuracy of the Naïve Bayes Method using data testing is 92.0502%. With a description of the data stock price 149 classified experience raising prices, and there are 6 data stock prices are supposed to experience raising stock prices but classified stock price decline. And there are 71 data stock prices that are classified correctly stock price decline, but there are 13 stock price data that should experience a decline in the stock price, but classified experience ascent stock price. with the presentation for the value of precision or accuracy of information expected by the author with the answers given by the system of 0.920 and the value for the recall or the success rate of taking action against information found by the system of the prediction results While 0.961 using Decision Tree J-48 results for accuracy using data testing is 98.7448%. %. With 154 data stock price information that is classified is experiencing ascent rates, and there is 1 data stock prices are supposed to experience the ascent rates, however classified stock price decline. And there is a stock price data 82 classified properly stock price decline, but there are 2 stock price data that should experience a decline in the stock price, but classified experience ascent stock price. With the presentation for the value of precision or accuracy of information expected by the author with the answers given by the system of 0.989 and value for the recall or the success rate of taking action against information found by the system of 0.997 Keywords: Stock price, Indonesia stock exchange, Naїve Bayes, Decision tree-J48, WEKA.