Fuzzy forecasting based on fuzzy-trend logical relationship groups

Area of Science:

• Computational intelligence
• Financial forecasting
• Time series analysis

Background:

• Accurate stock market prediction is crucial for financial decision-making.
• Existing forecasting methods often struggle with the inherent complexity and volatility of stock market data.
• The Taiwan Stock Exchange Capitalization Weighted Stock Index (TAIEX) presents a challenging forecasting target due to its dynamic nature.

Purpose of the Study:

• To develop and validate a novel forecasting method for the TAIEX using fuzzy-trend logical relationship groups (FTLRGs).
• To enhance the accuracy of stock market index prediction by leveraging fuzzy logic and trend analysis.
• To demonstrate the versatility of the proposed method by applying it to enrollment and inventory demand forecasting.

Main Methods:

• Clustering historical data into variable-length intervals using an automatic clustering algorithm.
• Defining fuzzy sets based on these intervals to represent data trends.
• Fuzzifying historical data to derive fuzzy logical relationships.
• Grouping fuzzy logical relationships into fuzzy-trend logical relationship groups (FTLRGs) for prediction.

Main Results:

• The proposed FTLRG method achieved higher average forecasting accuracy rates compared to existing methods in TAIEX prediction.
• The method also showed promising results when applied to forecasting enrollments and inventory demand.
• The experimental validation confirmed the effectiveness and robustness of the FTLRG approach.

Conclusions:

• The FTLRG method offers a significant improvement in forecasting accuracy for financial time series like the TAIEX.
• The approach is adaptable and effective for predicting other types of time-dependent data, such as enrollments and inventory demand.
• This research provides a valuable new tool for quantitative finance and demand forecasting.