The traditional factors of the clustering mutual fund (such as Net Asset Value (NAV)) are not always an efficient measure in both maximizing returns and minimizing portfolio risk. This research presents a novel measure, Net Asset Value Change Ratios for some of time durations N (NAVCR-N), to assist the mutual fund clustering. We proved the usage of the NAVCR-N as mutual fund LTF similarity measures and LTF are then selected from differing clusters to create a diversified mutual fund portfolio. Approximately a hundred mutual fund data different times from the set for the fiscal year 2010-2018 are applied in the experiment to evaluate the effectiveness of the random approach and our diverse approaches.
