The study of investment portfolio using K-means

Abstract

Nowadays, wealth management is a hot-term. Selecting a portfolio is one of the important steps. Many people, for example fund managers or investors, are willing to spend much time and resource on searching undervalue stocks, building different portfolio to achieve their investment target. The main idea of this project is to investigate the properties of different stocks and stock portfolio in Hong Kong stock market using K-means. K-means is one of the cluster analyses. It can group data objects (stock or portfolio) based on their information found in the data that describe the objects and their relationships. The objects within a group are similar to one another and different from the objects in other group. K-means is often exploratory in nature and frequently required postprocessing techniques to validate and explain the result. In this project, the data consists of 1081 listed stock in Hong Kong stock market and data was taken from 2004-1-1 to 2008-12-31. The simulation is implemented Java programming and MYSQL database.