PERAMALAN HARGA SAHAM PT UNILEVER INDONESIA MENGGUNAKAN FUZZY TIME SERIES MARKOV CHAIN ABSOLUTE DIFFERENCES K-MEANS CLUSTERING DAN FUZZY TIME SERIES MARKOV CHAIN TREE PARTITION METHOD

Stocks are an investment instrument that is in high demand and can provide large profits for investors, especially stocks from large and popular companies, such as PT Unilever Indonesia. However, the boycott action causes its share price to tend to decline and makes it difficult for investors to make investment decisions. In this research, the development of the Fuzzy Time Series Markov Chain (FTSMC) method, namely Fuzzy Time Series Markov Chain Absolute Differences K-Means Clustering (FTSMC-ADKC) and Fuzzy Time Series Markov Chain Tree Partition Method (FTSMC-TPM) is used in forecasting the stock price of PT Unilever Indonesia. Both methods have a good ability to forecast volatile time series data and do not require data distribution assumptions such as classical forecasting methods. Furthermore, the accuracy of the two methods is compared through the Mean Absolute Percentage Error (MAPE) value, resulting in FTSMC-ADKC being more accurate with a MAPE of 0.71% compared to FTSMC-TPM with a MAPE of 1.21%. By using FTSMC-ADKC, the closing price of PT Unilever Indonesia shares on March 27, 2024 is estimated to be Rp2724.46, which has decreased from the previous day.