Stock prediction and mutual fund portfolio management using curve fitting techniques

Area of Science:

• Quantitative Finance
• Data Mining
• Investment Management

Background:

• Share market investments offer higher profit potential but involve significant market risk.
• Risk aversion leads many investors to mutual funds managed by professionals.
• Mutual funds require complex calculations for stock price forecasting and risk management.

Purpose of the Study:

• To develop a data mining-based framework for diversifying mutual fund investments.
• To mitigate market risks associated with direct share market investments.
• To enhance investment yields for mutual fund investors.

Main Methods:

• Employed curve fitting/regression techniques for individual stock price forecasting.
• Proposed a capital diversification framework using a buy-and-hold strategy.
• Integrated statistical features and domain knowledge for investment decisions.

Main Results:

• The framework effectively diversifies capital sector-wise and company-wise.
• Achieved higher returns with lower risks compared to benchmarks.
• Demonstrated strong performance in the Indian stock market.

Conclusions:

• The proposed framework offers a robust approach to mutual fund investment diversification.
• Data mining and strategic allocation can improve investment outcomes.
• The method provides a viable strategy for balancing risk and return in equity investments.