Anongnart Srivihok

Comparisons of classifier algorithms: Bayesian network, C4.5, decision forest and NBTree for Course Registration Planning model of undergraduate students

The success rate of computer science and engineering students in private universities are not high. It is helpful to find the model to assist students in registration planning. The objective of this research is to propose the classifier algorithm for building course registration planning model (CRPM) from historical dataset. The algorithm is selected by comparing performances of four classifiers include Bayesian network, C4.5, Decision Forest and NBTree. The dataset were obtained from student enrollments including grade point average (GPA) and grades of undergraduate students whose majors were computer science or computer engineering. These dataset included grades in each subject of first and second year students from a private university in Thailand. Results showed that NBTree seemed to be the best of four classifiers which had highest prediction power. NBTree was used to generate CRP model which can be used to predict student class of GPA and consider student course sequences for registration planning.

Comparison of attribute selection techniques and algorithms in classifying bad behaviors of vocational education students

This study presents the comparison of attribute selection techniques which used for classifying the bad behaviors of vocational education students. There are two classification methods: hybrid classification and single classification. Hybrid classification includes two steps, step one is attribute selection by search method using genetic search and results are compared by three evaluators: 1) Correlation-based Feature Selection (CFS) 2) Consistency-based Subset Evaluation and 3) Wrapper Subset Evaluation. Step two is the classification of data set by using selected attributed from step one and four classification algorithms. Next, Simple classification used classification algorithms only without attribute selection. The four classification algorithms that used in this experiment for comparing in two methods are : 1) Naive Bayes classifier 2) Baysian Belief Network 3) C4.5 algorithm and 4) RIPPER algorithm. The measurements of classification efficiency had been obtained by using the k-fold cross validation technique. From the experiment, it was found that hybrid classification technique using genetic search and CFS evaluator with C4.5 algorithm, gives the highest accuracy rate at 82.52%. However, results from F-measure evaluation showed that C4.5 algorithm did not fit for all data types. The hybrid classification technique using genetic search and wrapper subset with Baysian belief network can give a better precision value which can be seen in the F-measure, and it gives the accuracy rate at 82.42%.

An evaluation model for e-learning Websites in Thailand University

The objective of this research is to evaluate e-learning Websites for university in Thailand by using quality criteria which based on IEEE 1061 and ISO/IEC 9126. In this research, we determine 6 quality aspects (include 2 new aspects) such as functionality, reliability, usability, efficiency, maintainability and portability. In the research process, we categorize subjects into 5 groups. The result has found that the average of quality in e-learning Websites is 50.34% and this can be exploited as beneficial information for e-learning Webmaster to evaluate and improve the quality of e-learning Website according to the proposed quality model for optimizing the effective e-learning.

Personalized Trip Information for E-Tourism Recommendation System Based on Bayes Theorem

This paper presents the personalized recommendation system for e-tourism by using statistic technique base on Bayes Theorem to analyze user behaviors and recommend trips to specific users. The system is evaluated by using Recall, Precision and F-measure. Results demonstrate that it is possible to develop Personalization Recommendation System. Past and recent information of customer’s behaviors are used for recommending products and services best fit to relevant customers in the e-Tourism.

Preprocessing of imbalanced breast cancer data using feature selection combined with over-sampling technique for classification

Class imbalance problems have been found in many medical data in recent years. Data are imbalanced when the distributions of classes are highly imbalanced that means the number of instances of one class is very different to the other classes. Feature selection combined with over-sampling technique (FOT) is proposed to preprocess data before classifying our dataset, imbalanced breast cancer. We used feature selection techniques, Consistency Subset Evaluation, at the beginning to remove insignificant attributes of data. The remaining attributes were fed into over-sampling phase to adjust instances in the minority class. After preprocessed the dataset, we classified data using three classification algorithms, decision tree, BayesNet, and OneR. The f-values of classification data using FOT are 0.76, 0.638, and 0.64, respectively. These are greater than the f-values of three above classifications without FOT as, 0.561, 0.518, and 0.512, respectively. The experimental results indicated that FOT achieves better f-values than non-FOT preprocessing and have performed well in improving the performance of classifiers on this dataset, especially, decision tree.

Using Bayesian Network for planning course registration model for undergraduate students

The objective of this study is to propose a model for planning course registration by using a data mining technique: Bayesian network. The proposed model can be used to predict the sequences of courses to be registered by undergraduate students whose majors are computer science or engineering. The data set was obtained from student enrollments and include GPA and grades in each subject for first and second year students from a private university in Thailand. Evaluations show that the predictive power of this model is acceptable. The implications from this studypsilas findings suggest that the model can be applied for advising students in planning courses to be registered in each semester. Further, the model appears to be useful for improving curriculum development in order to fit both studentspsila and university requirements.

Inbound tourists segmentation with combined algorithms using K-Means and Decision Tree

Tourism is one of the main industries which bring about monetary to its country. To survive in the competitive industries these tourism organizations must have innovative strategies to carry on their business. One of the tools is tourism market segmentation which is used for strategic planning. This study presents inbound tourist market segmentation with combined algorithms using K-Means and Decision Tree. The study was divided into two phases. In the clustering phase, the segmentation was performed by Self Organizing Map (SOM) and K-Means. SOM used for determining the appropriate number of cluster. Then, K-Means used for refined the tourist clusters. The results of clustering phase were analyzed. In the classification phase, three classifiers were compared the performances of predictability by using the output provided by K-Means, i.e. Decision Tree, NaYve Bayes and Multilayer Perceptron (MLP). The experimental results indicated that SOM provided 6 clusters and K-Means gave better performance than SOM guided by Silhouette, Root Means Square Standard Deviation (RMSSTD) and R Square (RS). The predictive ability of J48 Decision Tree outperformed both of MLP and NaYve Bayes based on the tourist variables. J48 Decision Tree indicated the accuracy as 99.54%. The results of this study can be used for tourism management products and services.

Duo Bundling Algorithms for Data Preprocessing: Case Study of Breast Cancer Data Prediction

 Abstract—Classification of imbalanced dataset is the most popular and challenged problems for researchers to solve in nowadays. This paper proposed a two-steps approach to improve the quality of class prediction imbalanced breast cancer dataset. The two-steps approach consists of two main techniques: 1) using feature selection techniques to filter out unimportant features from the dataset; and 2) using the over-sampling technique to adjust the size of the minority class to be similar to the size of the majority class. The three different classification algorithms: artificial neural network (MLP), decision tree (C4.5) and Nai ve Bayes, were applied. The classification result indicated that C4.5 was the most suitable to classify this dataset which can give the highest accuracy of 83.80%.

Association rule mining for project performance measurement of Technology Transfer Centers in Thailand

In Thailand, there are many Technology Transfer Centers (TTC) located in every district of the country. However, the evaluation of their performance is difficult and complex since TTCs are different in size, location, type of training and expertise. Accordingly, the study of TTC project performance in Thailand is scanty due to the limitations of confidentiality. This study uses data mining techniques to analyze historical data of project profiles from Technology Transfer Centers (TTC) in Thailand. TTC data are segmented into groups according to their performance. This method is used to separate value centers from the inactive ones. Then a priori algorithm is applied to find the relationships within TTC features and project profiles of clusters having different performances. The results of this study will benefit the organization in exploiting it to bring about a competitive advantage and being able to improve the efficiency of TTC as well as its application to strategic planning.

Comparative Feature Selection of Crime Data in Thailand

The crime is a major problem of community and society which is increasing day by day. Especially in Thailand, crime is a major problem that affects all aspects of the country such as tourism, administration of government and problem in daily life. Therefore, government and private sectors have to understand the several crime patterns for planning, preventing and solving solution of crime correctly. The purposes of this study are to generate a crime model for Thailand using data mining techniques. Data were collected from Dailynews and Thairath online newspapers. The proposed model can be generated by using more feature selection and more classification techniques to different model. Experiments show feature selection with the wrapper of attribute evaluator seems to be an appropriate evaluation algorithm because data set mostly is the best accuracy rate. This improves efficiency in identifying offenders more quickly and accurately. The model can be used for the prevention of crime that will occur in Thailand in the future.