Mutual Fund NAV Prediction Using Cascaded SVM Models

The world of finance is often viewed as a gamble because of its unpredictability. Although NAVs are not as volatile as stock prices, the mutual fund market too has its share of uncertainties. The current models for the prediction of NAVs are based on historical NAV data using neural network models or regression models. The paper aims to predict NAV by expanding the focus onto various other parameters as input, including macroeconomic factors. In addition, results have been optimized using SVM in a cascaded form. The experiments were conducted for the Indian market, in specific, the HDFC mid cap opportunities (Growth) mutual fund scheme.