Jr-Shian Chen

Fuzzy time series model based on probabilistic approach and rough set rule induction for empirical research in stock markets

This study proposes a hybrid fuzzy time series model with two advanced methods, cumulative probability distribution approach (CPDA) and rough set rule induction, to forecast stock markets. To improve forecasting accuracy, three refining processes of fuzzy time series are provided in the proposed model: (1) using CPDA to discretize the observations in training datasets based on the characteristics of data distribution, (2) generating rules (fuzzy logical relationships) by rough set algorithm and (3) producing forecasting results based on rule support values from rough set algorithm. To verify the forecasting performance of the proposed model in detail, two empirical stock markets (TAIEX and NYSE) are used as evaluating databases; two other methodologies, proposed by Chen and Yu, are used as comparison models, and two different evaluation methods (moving windows) are used. The proposed model shows a greatly improved performance in stock market forecasting compared to other fuzzy time series models.

Extracting classification rule of software diagnosis using modified MEPA

Defective software modules cause software failures, increase development and maintenance costs, and reduce customer satisfaction. Effective defect prediction models can help developers focus quality assurance activities on defect-prone modules and thus improve software quality by using resources more efficiently. In real-world databases are highly susceptible to noisy, missing, and inconsistent data. Noise is a random error or variance in a measured variable [Han, J., & Kamber, M. (2001). Data Mining: Concepts and Techniques, San Francisco: Morgan Kaufmann Publishers]. When decision trees are built, many of the branches may reflect noisy or outlier data. Therefore, data preprocessing steps are very important. There are many methods for data preprocessing. Concept hierarchies are a form of data discretization that can use for data preprocessing. Data discretization has many advantages, such as data can be reduced and simplified. Using discrete features are usually more compact, shorter and more accurate than using continuous ones [Liu, H., Hussain, F., Tan, C.L., & Dash, M. (2002). Discretization: An enabling technique. Data Mining and Knowledge Discovery, 6(4), 393-423]. In this paper, we propose a modified minimize entropy principle approach and develop a modified MEPA system to partition the data, and then build the classification tree model. For verification, two NASA software projects KC2 and JM1 are applied to illustrate our proposed method. We establish a prototype system to discrete data from these projects. The error rate and number of rules show that the proposed approach is both better than other methods.

Forecasting tourism demand based on improved fuzzy time series model

The total tourist arrivals, is an important factor to understand the tourism market and to predict the trend of tourism demand, is necessity and exigency for tourism demand and hospitality industries for subsequent planning and policy marketing. This paper proposed a fusion model of fuzzy time-series to improve the forecasting accuracy on total tourist arrivals, which consider the cluster characteristic of observations, define more persuasive universe of discourse based on k-mean approach, fuzzify the observation precisely by triangular fuzzy number, establish fuzzy logical relationships groups by employing rough set rule induction, and assign weight to various fuzzy relationship based on rule-support. In empirical case study, the proposed model is verified by using tourist datasets and comparing forecasting accuracy with listed models. The experimental results indicate that the proposed approach outperforms listed models with lower mean absolute percentage error.

A learning curve of total knee arthroplasty (TKA) based on surgical volume analysis.

The purpose of this study is to discover valuable medical facts by utilizing the Taiwan National Health Insurance (NHI) database, which contains 32,200 records of TKA surgeries. Three main objectives of this paper include the following: (a) building learning curves of TKA from the target database; (b) characterizing how the TKA volume correlates with infection rate and mortality; (c) examining the differences of infection rate and mortality between the medical center (Group I) and the non-medical center (Group II). The TKA samples are classified into two groups according to their institution type (medical center and non-medical center). The Z-test is used to test whether there are differences in the infection rate and mortality between the two observed groups. This study also adopts linear/nonlinear regression to investigate the relationship between TKA volume and the infection rate (mortality). This study has three main findings: (a) it confirms a correlation between the TKA surgical volumes and certain outcomes, (b) surgeons and hospitals with higher TKA volumes exhibit better operation quality, lower postoperative complication rate, and (c) there are significant differences in infection and mortality rate between Group I and Group II.

Software diagnosis using fuzzified attribute base on modified MEPA

Currently, there are many data preprocess methods, such as data discretization, data cleaning, data integration and transformation, data reduction ... etc. Concept hierarchies are a form of data discretization that can use for data preprocessing. Using discrete data are usually more compact, shorter and more quickly than using continuous ones. So that we proposed a data discretization method, which is the modified minimize entropy principle approach to fuzzify attribute and then build the classification tree. For verification, two NASA software projects KC2 and JM1 are applied to illustrate our proposed method. We establish a prototype system to discrete data from these projects. The error rate and number of rules show that the proposed approaches are both better than other methods.

A new e-learning achievement evaluation model based on RBF-NN and similarity filter

As Internet rises fast in recent decades, teaching and learning tools based on Internet technology are rapidly applied in education. Learning through Internet can make learners absorb knowledge without the limitations on learning time and distance. Therefore, in academy, e-learning is one of the popular learning assistant instruments. Recently, “student-centered” instruction has become one of the primary approaches in education, and the e-learning system, which can provide the learning environment of personalization and adaptability, is more and more popular. By using e-learning system, teachers can adjust the learning schedule instantly for learners according to their learning achievements, and build more adaptive learning environments. However, in some cases, bias assessments are given for student achievements under specific uncontrollable conditions (i.e. tiredness, preference). In dire need of overcoming this predicament, a new model based on radial basis function neural networks (RBF-NN) and similarity filter to evaluate learning achievements is proposed. The proposed model includes three phases to reduce bias assessments: (1) preprocess: select important features (attributes) to enhance classification performance by feature selection methods and utilize minimal entropy principle approach (MEPA) to fuzzify the quantitative data, (2) similarity filter: select linguistic values for each feature and delete inconsistent data by the similarity threshold (similarity filter) and (3) construct classification model and accuracy evaluation: build the proposed model based on RBF-NN and evaluate model performance. To verify the proposed model, a practical achievement dataset, collected from e-learning online examination system in a university of Taiwan, is used as experiment dataset, and the performance of the proposed model is compared with the listing models in this paper. From the empirical study, it is shown that the proposed model provided more proper achievement evaluations than the listing models.

Constructing New Drug Adoption Rules in Medical Institution

Knowledge of medicinal chemical compounds is complex and evidence-based. This study extracted knowledge of drugs from experts and assessed that knowledge using sophisticated computational techniques for critical decision making. By using fuzzy inference system this study evaluated new drug candidates in efficient way by considering not only safety and efficacy, but also the hottest issue-cost. The process utilized in this study was based on the consensus of the Drugs and Therapeutic Committee (DTC). Finally, this study establishes new drug adoption rules based on the knowledge of DTC members and implements a fuzzy inference system. Notably, the verification and comparison of this study also provide an intuitive direction for system development and further researches.

An Alternate Method Of Examining Ipo Returns

Abstract The study proposes an alternate method, a soft computing system that mixes an entropy-based discretization method, fuzzy-rule-similarity, and rough set theory, for solving the real-world problems of initial public offerings (IPO) returns faced by both academicians and practitioners. The proposed method is illustrated by examining two practiced datasets for publicly traded firms in Taiwan. The experimental results reveal that the proposed method outperforms the listing methods in terms of accuracy and number of rules. Furthermore, the proposed method generates a set of comprehensible decision rules that can be applied in a knowledge-based system for investment decision-making of investors.

DIAGNOSING CARDIOVASCULAR DISEASE USING AN ENHANCED ROUGH SETS APPROACH

Cardiovascular disease is a chronic disease and an ongoing threat to human health. Clinical data, including chemistry analysis data and electrocardiogram (ECG) data for heartbeat behavior, are commonly used to classify the cardiovascular diseases in supporting medical diagnosis. This study proposes a new approach for enhancing rough set classifier which applied to diagnose cardiovascular disease. Two datasets were used in this empirical case study to illustrate the proposed approach. Due to its improved accuracy and fewer rules, the proposed approach is superior to listing methods.

Analyzing total knee arthroplasty using surgical volume of different hospital levels

The purpose of this research is to discover valuable medical facts by mining TKA (total knee arthroplasty) surgical volume of three different hospital levels from Taiwan NHI (National Health Insurance) database. In this paper, there are three main objectives provided: (1) to build up the learning curves using the patients¿ outcomes of three different hospital levels; (2) to characterize whether TKA surgical volume correlates with infection rate and mortality rate; and (3) to examine whether there are differences in the infection rate and mortality rate pair-wise comparisons among medical center, district hospital, and local hospital. After building up the learning curves, the results of Z-test confirm the differences between these three observed levels. For verification, we¿ve interviewed five specialist surgeons with a semi-structure questionnaire containing some close questions and some open-end, and the investigative results are consistent with the results from data analysis.