Applying Fuzzy Candlestick Pattern Ontology to Investment Knowledge Management

Abstract

A fuzzy candlestick pattern based ontology is proposed for assisting candlestick pattern representation, storage, and reuse. Japanese candlestick theory is a widely used technical analysis method for stock and commodity investment decision making. The theory assumes that candlestick patterns reflect the psychology of market, and investors make investment decision by observing the pattern in the Candlestick chart. A candlestick pattern is composed of candlestick lines. We model the different part of a candlestick line with fuzzy linguistic variables and transfer the financial time series data to fuzzy candlestick lines. The user can use data mining algorithm such as decision tree to mine some candlestick patterns for investment decision making and the mined candlestick patterns could be stored in a database for different user's future reuse. Based on the proposed approach, we implement a system prototype to get experimental results. Our approach can be future used with other financial time series prediction results to provide users more information for investment decision making.