Using Z-Score for Mixed-Fund Clustering

For financial planners, preparing data to provide advice to investors is crucial. Mutual funds are one of the investment options managed by fund managers. So, grouping mutual funds as supporting information for investor’s decision-making can help them make better choices. However, the data for analysis of mutual funds required is extensive, and the analysis needs to be conducted accurately, continuously, and rapidly. This project aims to group mutual funds with mixed factors based on return rates and to develop a program for grouping mutual funds using k-means clustering. The study uses daily net asset value data of mutual funds from January 1, 2021, to June 30, 2024. Grouping method is performed by using Z-Score of the expected return and standard deviation for 1-year and 3-year periods, in conjunction with considering the weights of the factors using RapidMiner. The python program which requires only input data on the study period and mutual fund type is also developed. The findings show that, based on this dataset, mixed mutual funds can be grouped into three clusters. The program can accurately group the funds according to the defined steps and methods, thereby it improves efficiency and reduces errors compared to manual grouping.