Data mining technology in mutual funds evaluation

Ben Chang Shia, Guo Xiang Xu, Lung Hua Sung, Ya Wen Lin, Hsiao Ho Lin

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

There are many Mutual Funds rating companies in the market. Rating companies have different standards to evaluate Mutual Funds such as risks, incomes and some referable indexes. The study is aim at setting a score criterion to discuss the performance of equity fund, ETF and Index Fund. The objects of this study include eight domestic funds which are set up more than one year. We collect the data which are set up before August, 2008. The study uses Text Mining to choose proper variables. By Cluster Analysis, the similar characteristics of risk and return can be distributed into same group. We analyze the performance of the Mutual Funds for each group. Furthermore, each fund gets scores through the score criterion and tries to compare the study with Morningstar, Lipper, and Fund-Watch. Finally, Discriminant Analysis can determine that new data may be distributed into which cluster. The main conclusions are: 1. According to risk and return, Cluster Analysis can divide Large-sized Equity Fund into three clusters. ETF and Index Fund can also divide into three clusters by the same way. Besides, there are obvious differences between each cluster. 2. On the basis of the result, JF (TAIWAN) Taiwan Fund is the best performance in Large-sized Equity Fund. Polaris Taiwan Top 50 Tracker Fund is the best in ETF and Index Fund. 3. The accuracy is about 96.03% in Large-sized Equity Funds and 97.35% in Index Funds by Discriminant Analysis. In term of the study, it is immediate to find out the result of cluster and performance evaluation for new data. 4. Morningstar, Lipper and Fund-Watch are common rating companies in the market, but the rating of Large-sized Equity Fund is different. In business recession, they can evaluate a better fund relatively, but the evaluated fund can't increase profit. In this study, the way of rating bases on the outcome of Cluster Analysis and score of the funds. It can not only reduce the loss and risk but also provide objective and useful information to investors.

Original languageEnglish
Title of host publicationProceedings - 2009 International Conference on New Trends in Information and Service Science, NISS 2009
Pages896-903
Number of pages8
DOIs
Publication statusPublished - 2009
Externally publishedYes
Event2009 International Conference on New Trends in Information and Service Science, NISS 2009 - Beijing, China
Duration: Jun 30 2009Jul 2 2009

Other

Other2009 International Conference on New Trends in Information and Service Science, NISS 2009
CountryChina
CityBeijing
Period6/30/097/2/09

Fingerprint

Data mining
Cluster analysis
Watches
Discriminant analysis
Industry
Profitability

ASJC Scopus subject areas

  • Computer Science Applications
  • Software

Cite this

Shia, B. C., Xu, G. X., Sung, L. H., Lin, Y. W., & Lin, H. H. (2009). Data mining technology in mutual funds evaluation. In Proceedings - 2009 International Conference on New Trends in Information and Service Science, NISS 2009 (pp. 896-903). [5260780] https://doi.org/10.1109/NISS.2009.237

Data mining technology in mutual funds evaluation. / Shia, Ben Chang; Xu, Guo Xiang; Sung, Lung Hua; Lin, Ya Wen; Lin, Hsiao Ho.

Proceedings - 2009 International Conference on New Trends in Information and Service Science, NISS 2009. 2009. p. 896-903 5260780.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Shia, BC, Xu, GX, Sung, LH, Lin, YW & Lin, HH 2009, Data mining technology in mutual funds evaluation. in Proceedings - 2009 International Conference on New Trends in Information and Service Science, NISS 2009., 5260780, pp. 896-903, 2009 International Conference on New Trends in Information and Service Science, NISS 2009, Beijing, China, 6/30/09. https://doi.org/10.1109/NISS.2009.237
Shia BC, Xu GX, Sung LH, Lin YW, Lin HH. Data mining technology in mutual funds evaluation. In Proceedings - 2009 International Conference on New Trends in Information and Service Science, NISS 2009. 2009. p. 896-903. 5260780 https://doi.org/10.1109/NISS.2009.237
Shia, Ben Chang ; Xu, Guo Xiang ; Sung, Lung Hua ; Lin, Ya Wen ; Lin, Hsiao Ho. / Data mining technology in mutual funds evaluation. Proceedings - 2009 International Conference on New Trends in Information and Service Science, NISS 2009. 2009. pp. 896-903
@inproceedings{412ce6e6cc61456786d5ce1907f25121,
title = "Data mining technology in mutual funds evaluation",
abstract = "There are many Mutual Funds rating companies in the market. Rating companies have different standards to evaluate Mutual Funds such as risks, incomes and some referable indexes. The study is aim at setting a score criterion to discuss the performance of equity fund, ETF and Index Fund. The objects of this study include eight domestic funds which are set up more than one year. We collect the data which are set up before August, 2008. The study uses Text Mining to choose proper variables. By Cluster Analysis, the similar characteristics of risk and return can be distributed into same group. We analyze the performance of the Mutual Funds for each group. Furthermore, each fund gets scores through the score criterion and tries to compare the study with Morningstar, Lipper, and Fund-Watch. Finally, Discriminant Analysis can determine that new data may be distributed into which cluster. The main conclusions are: 1. According to risk and return, Cluster Analysis can divide Large-sized Equity Fund into three clusters. ETF and Index Fund can also divide into three clusters by the same way. Besides, there are obvious differences between each cluster. 2. On the basis of the result, JF (TAIWAN) Taiwan Fund is the best performance in Large-sized Equity Fund. Polaris Taiwan Top 50 Tracker Fund is the best in ETF and Index Fund. 3. The accuracy is about 96.03{\%} in Large-sized Equity Funds and 97.35{\%} in Index Funds by Discriminant Analysis. In term of the study, it is immediate to find out the result of cluster and performance evaluation for new data. 4. Morningstar, Lipper and Fund-Watch are common rating companies in the market, but the rating of Large-sized Equity Fund is different. In business recession, they can evaluate a better fund relatively, but the evaluated fund can't increase profit. In this study, the way of rating bases on the outcome of Cluster Analysis and score of the funds. It can not only reduce the loss and risk but also provide objective and useful information to investors.",
author = "Shia, {Ben Chang} and Xu, {Guo Xiang} and Sung, {Lung Hua} and Lin, {Ya Wen} and Lin, {Hsiao Ho}",
year = "2009",
doi = "10.1109/NISS.2009.237",
language = "English",
isbn = "9780769536873",
pages = "896--903",
booktitle = "Proceedings - 2009 International Conference on New Trends in Information and Service Science, NISS 2009",

}

TY - GEN

T1 - Data mining technology in mutual funds evaluation

AU - Shia, Ben Chang

AU - Xu, Guo Xiang

AU - Sung, Lung Hua

AU - Lin, Ya Wen

AU - Lin, Hsiao Ho

PY - 2009

Y1 - 2009

N2 - There are many Mutual Funds rating companies in the market. Rating companies have different standards to evaluate Mutual Funds such as risks, incomes and some referable indexes. The study is aim at setting a score criterion to discuss the performance of equity fund, ETF and Index Fund. The objects of this study include eight domestic funds which are set up more than one year. We collect the data which are set up before August, 2008. The study uses Text Mining to choose proper variables. By Cluster Analysis, the similar characteristics of risk and return can be distributed into same group. We analyze the performance of the Mutual Funds for each group. Furthermore, each fund gets scores through the score criterion and tries to compare the study with Morningstar, Lipper, and Fund-Watch. Finally, Discriminant Analysis can determine that new data may be distributed into which cluster. The main conclusions are: 1. According to risk and return, Cluster Analysis can divide Large-sized Equity Fund into three clusters. ETF and Index Fund can also divide into three clusters by the same way. Besides, there are obvious differences between each cluster. 2. On the basis of the result, JF (TAIWAN) Taiwan Fund is the best performance in Large-sized Equity Fund. Polaris Taiwan Top 50 Tracker Fund is the best in ETF and Index Fund. 3. The accuracy is about 96.03% in Large-sized Equity Funds and 97.35% in Index Funds by Discriminant Analysis. In term of the study, it is immediate to find out the result of cluster and performance evaluation for new data. 4. Morningstar, Lipper and Fund-Watch are common rating companies in the market, but the rating of Large-sized Equity Fund is different. In business recession, they can evaluate a better fund relatively, but the evaluated fund can't increase profit. In this study, the way of rating bases on the outcome of Cluster Analysis and score of the funds. It can not only reduce the loss and risk but also provide objective and useful information to investors.

AB - There are many Mutual Funds rating companies in the market. Rating companies have different standards to evaluate Mutual Funds such as risks, incomes and some referable indexes. The study is aim at setting a score criterion to discuss the performance of equity fund, ETF and Index Fund. The objects of this study include eight domestic funds which are set up more than one year. We collect the data which are set up before August, 2008. The study uses Text Mining to choose proper variables. By Cluster Analysis, the similar characteristics of risk and return can be distributed into same group. We analyze the performance of the Mutual Funds for each group. Furthermore, each fund gets scores through the score criterion and tries to compare the study with Morningstar, Lipper, and Fund-Watch. Finally, Discriminant Analysis can determine that new data may be distributed into which cluster. The main conclusions are: 1. According to risk and return, Cluster Analysis can divide Large-sized Equity Fund into three clusters. ETF and Index Fund can also divide into three clusters by the same way. Besides, there are obvious differences between each cluster. 2. On the basis of the result, JF (TAIWAN) Taiwan Fund is the best performance in Large-sized Equity Fund. Polaris Taiwan Top 50 Tracker Fund is the best in ETF and Index Fund. 3. The accuracy is about 96.03% in Large-sized Equity Funds and 97.35% in Index Funds by Discriminant Analysis. In term of the study, it is immediate to find out the result of cluster and performance evaluation for new data. 4. Morningstar, Lipper and Fund-Watch are common rating companies in the market, but the rating of Large-sized Equity Fund is different. In business recession, they can evaluate a better fund relatively, but the evaluated fund can't increase profit. In this study, the way of rating bases on the outcome of Cluster Analysis and score of the funds. It can not only reduce the loss and risk but also provide objective and useful information to investors.

UR - http://www.scopus.com/inward/record.url?scp=70449604903&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=70449604903&partnerID=8YFLogxK

U2 - 10.1109/NISS.2009.237

DO - 10.1109/NISS.2009.237

M3 - Conference contribution

SN - 9780769536873

SP - 896

EP - 903

BT - Proceedings - 2009 International Conference on New Trends in Information and Service Science, NISS 2009

ER -