Catálogo de publicaciones - libros

Text classification can greatly improve the performance of information retrieval and information filtering, but high dimensionality of documents baffles the applications of most classification approaches. This paper proposed a Difference-Similitude Matrix (DSM) based method to solve the problem. The method represents a pre-classified collection as an item-document matrix, in which documents in same categories are described with similarities while documents in different categories with differences. Using the DSM reduction algorithm, simpler and more efficient than rough set reduction, we reduced the dimensionality of document space and generated rules for text classification.

The increasing amount and complexity of data used in predictive toxicology calls for efficient and effective feature selection methods in data pre-processing for data mining. In this paper, we propose a kNN model-based feature selection method (kNNMFS) aimed at overcoming the weaknesses of ReliefF method. It modifies the ReliefF method by: (1) using a kNN model as the starter selection aimed at choosing a set of more meaningful representatives to replace the original data for feature selection; (2) integration of the Heterogeneous Value Difference Metric to handle heterogeneous applications – those with both ordinal and nominal features; and (3) presenting a simple method of difference function calculation. The performance of kNNMFS was evaluated on a toxicity data set Phenols using a linear regression algorithm. Experimental results indicate that kNNMFS has a significant improvement in the classification accuracy for the trial data set.

Feature selection is effective in removing irrelevant data. However, the result of feature selection in unsupervised learning is not as satisfying as that in supervised learning. In this paper, we propose a novel methodology ULAC (Feature Selection for Unsupervised Learning Based on Attribute Correlation Analysis and Clustering Algorithm) to identify important features for unsupervised learning. We also apply ULAC into prosecutors’ office to solve the real world application for unsupervised learning.

Fish-net algorithm is a novel field learning algorithm which derives classification rules by looking at the range of values of each attribute instead of the individual point values. In this paper, we present a Feature Selection Fish-net learning algorithm to solve the Dual Imbalance problem on text classification. Dual imbalance includes the instance imbalance and feature imbalance. The instance imbalance is caused by the unevenly distributed classes and feature imbalance is due to the different document length. The proposed approach consists of two phases: (1) select a feature subset which consists of the features that are more supportive to difficult minority class; (2) construct classification rules based on the original Fish-net algorithm. Our experimental results on Reuters21578 show that the proposed approach achieves better balanced accuracy rate on both majority and minority class than Naive Bayes MultiNomial and SVM.

Most NDR algorithms need to solve large-scale eigenvalue problems or some variation of eigenvalue problems, which is of quadratic complexity of time and might be unpractical in case of large-size data sets. Besides, current algorithms are global, which are often sensitive to noise and disturbed by ill-conditioned matrix. In this paper, we propose a novel self-organizing NDR algorithm: SIE. The time complexity of SIE is O(NlogN). The main computing procedure of SIE is local, which improves the robustness of the algorithm remarkably.

Recently, Biometrics-based personal identification is regarded as an effective method of person’s identity with recognition automation and high performance. In this paper, the palmprint recognition method based on Hu invariant moment is proposed. And the low-resolution (75dpi) palmprint image (135×135 Pixel) is used for the small scale database of the effectual palmprint recognition system. The proposed system is consists of two parts: firstly, the palmprint fixed equipment for the acquisition of the correctly palmprint image and secondly, the algorithm of the efficient processing for the palmprint recognition.

In this paper, we extend the locally nearest neighbor classifiers to tackle the nonlinear classification problems via the kernel trick. The better performance is confirmed by the handwritten zip code digits classification experiments on the US Postal Service (USPS) database.

Web prefetching technique is one of the primary solutions used to reduce Web access latency and improve the quality of service. This paper makes use of Zipf’s 1st law and Zipf’s 2nd law to model the Web objects’ popularity, where Zipf’s 1st law is employed to model the high frequency Web objects and 2nd law for the low frequency Web objects, and proposes a PPM prediction model based on Web objects’ popularity for Web prefetching. A performance evaluation of the model is presented using real server logs. Trace-driven simulation results show that not only the model is easily to be implemented, but also can achieve a high prediction precision at the cost of relative low storage complexity and network traffic.

Axial representation plays a significant role in character recognition. The strokes of a character may consist of two regions, i.e. singular and regular regions. Therefore, a method to extract the central axis of a character requires two different processes to compute the axis in theses two different regions. The major problem of most traditional algorithms is that the extracted central axis in the singular region may be distorted by artifacts and branches. To overcome this problem, the wavelet-based amendment processing technique is developed to link the primary axis, so that the central axis in the singular region can be produced. Combining with our previously developed method for computing the primary axis in the regular region, we develop a novel scheme of extracting the central axis of character based on the wavelet transform (WT). Experimental results show that the final axis obtained from the proposed scheme closely resembles the human perceptions. It is applicable to both binary image and gray-level image as well. The axis representation is robust against noise.

Classifying dynamic behavioural based events, for example human behaviour profile, is a non-trivial task. In this paper, we propose an AI-based criteria-matching with classification algorithm which can be used to classify preference based decision outcome. The proposed algorithm is mathematically justified and with more practical benefits than a conventional multivariate discriminant analysis algorithm which is widely used for prediction tasks. Real world (Singapore) diamond dataset test results revealed the practical usefulness of our proposed algorithm to diamond sellers in Singapore.