Clustering stock price time series data to generate stock trading recommendations: An empirical study

摘要: Abstract Predicting the stock market is considered to be a very difficult task due its non-linear and dynamic nature. Our proposed system designed in such way that even layman can use it. It reduces burden on user. The user's job give only recent closing prices of as input Recommender will instruct him when buy sell if it profitable or not share case do trading. Using soft computing based techniques more suitable for predicting trends where data chaotic large number. systems are capable extracting relevant information from sets by discovering hidden patterns data. Here regression trees used dimensionality reduction clustering done with help Self Organizing Maps (SOM). assist investors identify possible profit-making opportunities also developing better understanding how extract price