Machine Learning Approach for Predicting the Net Asset Value (NAV) of Mutual Funds Based on Portfolio Holdings

Mutual funds, a cornerstone of modern investment portfolios, offer investors a diversified and professionally managed approach to financial growth. These investment vehicles collect small amount from multiple investors and invests the collective investment further in diversified portfolio of stocks, bonds, or other securities. Mutual funds are characterized by their net asset value (NAV), which represents the per-share value of the fund's assets minus its liabilities. Investors benefit from the expertise of seasoned fund managers who make investment decisions based on their analysis of market trends and company performance. By providing an accessible and diversified investment option, mutual funds have become a popular choice for both novice and seasoned investors seeking to achieve their financial goals. Apart from that, the role of AI is increasing exponentially in various areas. Such AI based tools and algorithms are utilized by various MFs expert to identify the suitable investment portfolios. This paper proposes a technique to analyse the performance of the different stock's portfolios based on the past data. Combining the isolated analysis of different stocks portfolios may be utilized to predict the net asset value prediction (NAV) of MFs. The experiments carried out in this research work uses linear regression analysis to analyse the stock's performance at sub level of the MFs. A case study for the Axis Bluechip fund's is also carried out by using the proposed ML approach. The Yahoo Finance dataset is used for the case study. Previously, machine learning algorithms like linear regression (LR), decision tree regression (DTR), and multivariate regression (MR) are taken to predict and analyse the dataset. These algorithms calculate the NAV of the mutual fund with good accuracy and analyse the best result for the test dataset. The various models are being created from different portfolios to analyse the NAV of the mutual funds. This work proposed hierarchical method to perform the linear regression analysis for the NAV predicting. This work uses R-Square (R2) statistical measures to determine the goodness of models. The obtained R2 is 0.86, implying that the created models have good accuracy and performance on the dataset.