Hahn-Ming Lee

Personalized e-learning system using Item Response Theory

Personalized service is important on the Internet, especially in Web-based learning. Generally, most personalized systems consider learner preferences, interests, and browsing behaviors in providing personalized services. However, learner ability usually is neglected as an important factor in implementing personalization mechanisms. Besides, too many hyperlink structures in Web-based learning systems place a large information burden on learners. Consequently, in Web-based learning, disorientation (losing in hyperspace), cognitive overload, lack of an adaptive mechanism, and information overload are the main research issues. This study proposes a personalized e-learning system based on Item Response Theory (PEL-IRT) which considers both course material difficulty and learner ability to provide individual learning paths for learners. The item characteristic function proposed by Rasch with a single difficulty parameter is used to model the course materials. To obtain more precise estimation of learner ability, the maximum likelihood estimation (MLE) is applied to estimate learner ability based on explicit learner feedback. Moreover, to determine an appropriate level of difficulty parameter for the course materials, this study also proposes a collaborative voting approach for adjusting course material difficulty. Experiment results show that applying Item Response Theory (IRT) to Web-based learning can achieve personalized learning and help learners to learn more effectively and efficiently.

DroidMat: Android Malware Detection through Manifest and API Calls Tracing

Recently, the threat of Android malware is spreading rapidly, especially those repackaged Android malware. Although understanding Android malware using dynamic analysis can provide a comprehensive view, it is still subjected to high cost in environment deployment and manual efforts in investigation. In this study, we propose a static feature-based mechanism to provide a static analyst paradigm for detecting the Android malware. The mechanism considers the static information including permissions, deployment of components, Intent messages passing and API calls for characterizing the Android applications behavior. In order to recognize different intentions of Android malware, different kinds of clustering algorithms can be applied to enhance the malware modeling capability. Besides, we leverage the proposed mechanism and develop a system, called Droid Mat. First, the Droid Mat extracts the information (e.g., requested permissions, Intent messages passing, etc) from each applications manifest file, and regards components (Activity, Service, Receiver) as entry points drilling down for tracing API Calls related to permissions. Next, it applies K-means algorithm that enhances the malware modeling capability. The number of clusters are decided by Singular Value Decomposition (SVD) method on the low rank approximation. Finally, it uses kNN algorithm to classify the application as benign or malicious. The experiment result shows that the recall rate of our approach is better than one of well-known tool, Androguard, published in Black hat 2011, which focuses on Android malware analysis. In addition, Droid Mat is efficient since it takes only half of time than Androguard to predict 1738 apps as benign apps or Android malware.

An efficient fuzzy classifier with feature selection based on fuzzy entropy

This paper presents an efficient fuzzy classifier with the ability of feature selection based on a fuzzy entropy measure. Fuzzy entropy is employed to evaluate the information of pattern distribution in the pattern space. With this information, we can partition the pattern space into nonoverlapping decision regions for pattern classification. Since the decision regions do not overlap, both the complexity and computational load of the classifier are reduced and thus the training time and classification time are extremely short. Although the decision regions are partitioned into nonoverlapping subspaces, we can achieve good classification performance since the decision regions can be correctly determined via our proposed fuzzy entropy measure. In addition, we also investigate the use of fuzzy entropy to select relevant features. The feature selection procedure not only reduces the dimensionality of a problem but also discards noise-corrupted, redundant and unimportant features. Finally, we apply the proposed classifier to the Iris database and Wisconsin breast cancer database to evaluate the classification performance. Both of the results show that the proposed classifier can work well for the pattern classification application.

A self-organizing HCMAC neural-network classifier

This paper presents a self-organizing hierarchical cerebellar model arithmetic computer (HCMAC) neural-network classifier, which contains a self-organizing input space module and an HCMAC neural network. The conventional CMAC can be viewed as a basis function network (BFN) with supervised learning, and performs well in terms of its fast learning speed and local generalization capability for approximating nonlinear functions. However, the conventional CMAC has an enormous memory requirement for resolving high-dimensional classification problems, and its performance heavily depends on the approach of input space quantization. To solve these problems, this paper presents a novel supervised HCMAC neural network capable of resolving high-dimensional classification problems well. Also, in order to reduce what is often trial-and-error parameter searching for constructing memory allocation automatically, proposed herein is a self-organizing input space module that uses Shannons entropy measure and the golden-section search method to appropriately determine the input space quantization according to the various distributions of training data sets. Experimental results indicate that the self-organizing HCMAC indeed has a fast learning ability and low memory requirement. It is a better performing network than the conventional CMAC for resolving high-dimensional classification problems. Furthermore, the self-organizing HCMAC classifier has a better classification ability than other compared classifiers.

Author Name Disambiguation for Citations Using Topic and Web Correlation

Today, bibliographic digital libraries play an important role in helping members of academic community search for novel research. In particular, author disambiguation for citations is a major problem during the data integration and cleaning process, since author names are usually very ambiguous. For solving this problem, we proposed two kinds of correlations between citations, namely, Topic Correlationand Web Correlation, to exploit relationships between citations, in order to identify whether two citations with the same author name refer to the same individual.The topic correlation measures the similarity between research topics of two citations; while the Web correlation measures the number of co-occurrence in web pages. We employ a pair-wise grouping algorithm to group citations into clusters. The results of experiments show that the disambiguation accuracy has great improvement when using topic correlation and Web correlation, and Web correlation provides stronger evidences about the authors of citations.

Two novel feature selection approaches for web page classification

To help the growing qualitative and quantitative demands for information from the WWW, efficient automatic Web page classifiers are urgently needed. However, a classifier applied to the WWW faces a huge-scale dimensionality problem since it must handle millions of Web pages, tens of thousands of features, and hundreds of categories. When it comes to practical implementation, reducing the dimensionality is a critically important challenge. In this paper, we propose a fuzzy ranking analysis paradigm together with a novel relevance measure, discriminating power measure (DPM), to effectively reduce the input dimensionality from tens of thousands to a few hundred with zero rejection rate and small decrease in accuracy. The two-level promotion method based on fuzzy ranking analysis is proposed to improve the behavior of each relevance measure and combine those measures to produce a better evaluation of features. Additionally, the DPM measure has low computation cost and emphasizes on both positive and negative discriminating features. Also, it emphasizes classification in parallel order, rather than classification in serial order. In our experimental results, the fuzzy ranking analysis is useful for validating the uncertain behavior of each relevance measure. Moreover, the DPM reduces input dimensionality from 10,427 to 200 with zero rejection rate and with less than 5% decline (from 84.5% to 80.4%) in the test accuracy. Furthermore, to consider the impacts on classification accuracy for the proposed DPM, the experimental results of China Time and Reuter-21578 datasets have demonstrated that the DPM provides major benefit to promote document classification accuracy rate. The results also show that the DPM indeed can reduce both redundancy and noise features to set up a better classifier.

Learning efficiency improvement of back-propagation algorithm by error saturation prevention method

Abstract Back-propagation (BP) algorithm is currently the most widely used learning algorithm in artificial neural networks. With proper selection of feed-forward neural network architecture, it is capable of approximating most problems with high accuracy and generalization ability. However, the slow convergence is a serious problem when using this well-known BP learning algorithm in many applications. As a result, many researchers take effort to improve the learning efficiency of BP algorithm by various enhancements. In this research, we consider that the error saturation (ES) condition, which is caused by the use of gradient descent method, will greatly slow down the learning speed of BP algorithm. Thus, in this paper, we will analyze the causes of the ES condition in the output layer. An error saturation prevention (ESP) function is then proposed to prevent the nodes in the output layer from the ES condition. We also apply this method to the nodes in hidden layers to adjust the learning terms. By the proposed methods, we can not only improve the learning efficiency by the ES condition prevention but also maintain the semantic meaning of the energy function. Finally, some simulations are given to show the workings of our proposed method.

A Grey-Based Nearest Neighbor Approach for Missing Attribute Value Prediction

This paper proposes a grey-based nearest neighbor approach to predict accurately missing attribute values. First, grey relational analysis is employed to determine the nearest neighbors of an instance with missing attribute values. Accordingly, the known attribute values derived from these nearest neighbors are used to infer those missing values. Two datasets were used to demonstrate the performance of the proposed method. Experimental results show that our method outperforms both multiple imputation and mean substitution. Moreover, the proposed method was evaluated using five classification problems with incomplete data. Experimental results indicate that the accuracy of classification is maintained or even increased when the proposed method is applied for missing attribute value prediction.

Prediction and evolutionary information analysis of protein solvent accessibility using multiple linear regression

A multiple linear regression method was applied to predict real values of solvent accessibility from the sequence and evolutionary information. This method allowed us to obtain coefficients of regression and correlation between the occurrence of an amino‐acid residue at a specific target and its sequence neighbor positions on the one hand, and the solvent accessibility of that residue on the other. Our linear regression model based on sequence information and evolutionary models was found to predict residue accessibility with 18.9% and 16.2% mean absolute error respectively, which is better than or comparable to the best available methods. A correlation matrix for several neighbor positions to examine the role of evolutionary information at these positions has been developed and analyzed. As expected, the effective frequency of hydrophobic residues at target positions shows a strong negative correlation with solvent accessibility, whereas the reverse is true for charged and polar residues. The correlation of solvent accessibility with effective frequencies at neighboring positions falls abruptly with distance from target residues. Longer protein chains have been found to be more accurately predicted than their smaller counterparts. Proteins 2005.

A neural network classifier with disjunctive fuzzy information

This paper presents a neural network classifier that learns disjunctive fuzzy information in the feature space. This neural network consists of two types of nodes in the hidden layer. The prototype nodes and exemplar nodes represent cluster centroids and exceptions in the feature space, respectively. This classifier automatically generates and refines prototypes for distinct clusters in the feature space. The number and sizes of these prototypes are not restricted, so the prototypes will form near-optimal decision regions to meet the distribution of input patterns and classify as many input patterns as possible. Next, exemplars will be created and expanded to learn the patterns that cannot be classified by the prototypes. Such a training strategy can reduce the memory requirement and speed up the process of non-linear classification. In addition, on-line learning is supplied in this classifier and the computational load is lightened. The experimental results manifest that this model can reduce the number of hidden nodes by determining the appropriate number of prototype nodes.