Liu Huan
Email Address
dcsliuh@nus.edu.sg
32 results
Publication Search Results
Now showing 1 - 10 of 32
Publication Feature selection for classification(1997) Dash, M.; Liu, H.; INFORMATION SYSTEMS & COMPUTER SCIENCEFeature selection has been the focus of interest for quite some time and much work has been done. With the creation of huge databases and the consequent requirements for good machine learning techniques, new problems arise and novel approaches to feature selection are in demand. This survey is a comprehensive overview of many existing methods from the 1970's to the present. It identifies four steps of a typical feature selection method, and categorizes the different existing methods in terms of generation procedures and evaluation functions, and reveals hitherto unattempted combinations of generation procedures and evaluation functions. Representative methods are chosen from each category for detailed explanation and discussion via example. Benchmark datasets with different characteristics are used for comparative study. The strengths and weaknesses of different methods are explained. Guidelines for applying feature selection methods are given based on data types and domain characteristics. This survey identifies the future research areas in feature selection, introduces newcomers to this field, and paves the way for practitioners who search for suitable methods for solving domain-specific real-world applications. © 1997 Elsevier Science B. V.Publication A low cost polymer based piezo-actuated micropump for drug delivery system(2004) Liu, H.; Tay, A.A.O.; Wan, S.Y.M.; Lim, G.C.; MECHANICAL ENGINEERING; COMPUTER SCIENCEThis paper presents the design, realization and simulation of a novel polymer based check-valve micropump actuated by piezoelectric disc. Comparing with silicon substrate, polymer materials have such advantages as flexibility, chemical and biological compatibility, 3D fabrication possibility and low cost in materials and mass production. Laser micromachining technology and precision engineering techniques were used to fabricate the prototype with the dimension of φ18mmx5.2mm. Results of preliminary experiments on fusion bonding between polyimide and polycarbonate are also presented. Using DI water as the pumping medium, the presented micropump is expected to achieve self-priming, bubble tolerance and low power consumption and a flow rate of 30μl/min at the resonance frequency of 300Hz.Publication Rule mining with prior knowledge-a belief networks approach(2001) Zhou, Z.; Liu, H.; Li, S.Z.; Chua, C.S.; COMPUTER SCIENCESome existing data mining methods, such as classification trees, neural networks and association rules, have the drawbacks that the user's prior knowledge cannot be easily specified and incorporated into the knowledge discovery process, and the rules mined from databases lack quantitative analyses. In this paper, we propose a belief networks method for rule mining, which takes the advantage of belief networks as the directed acyclic graph language and their function for numerical representation of probabilistic dependencies among the variables in the database, so that it can overcome the drawbacks. Since belief networks provide a natural representation for capturing causal relationship among a set of variables, our proposed method can mine more general correlation rules which can capture the relationship of more than two attribute variables. The potential application of the proposed method is demonstrated through the detailed case studies on benchmark databases. © 2001-IOS Press.Publication Efficient rule induction from noisy data(1996) Liu, H.; INFORMATION SYSTEMS & COMPUTER SCIENCEThe single-pattern-based rule induction method is sensitive to the order of data. Naturally, it is not suitable for data with noise. This paper reports a new rule induction method of this kind that handles noise effectively. Experiments are conducted to show that rules generated are compact and accurate, and the method is efficient and handles noise effectively. Its computational complexity is also given for comparison with other methods and as a guide for future application of this method.Publication Dimensionality reduction via discretization(1996-02) Liu, H.; Setiono, R.; INFORMATION SYSTEMS & COMPUTER SCIENCEThe existence of numeric data and large numbers of records in a database present a challenging task in terms of explicit concepts extraction from the raw data. The paper introduces a method that reduces data vertically and horizontally, keeps the discriminating power of the original data, and paves the way for extracting concepts. The method is based on discretization (vertical reduction) and feature selection (horizontal reduction). The experimental results show that (a) the data can be effectively reduced by the proposed method; (b) the predictive accuracy of a classifier (C4.5) can be improved after data and dimensionality reduction; and (c) the classification rules learned are simpler.Publication Source power estimation in linear arrays(1994) Ko, C.C.; Liu, H.; ELECTRICAL ENGINEERING; INFORMATION SYSTEMS & COMPUTER SCIENCEAn unbiased estimator was developed for the estimation of powers of uncorrelated directional sources through the use of linear array of sensors. Its performance was analyzed in depth, and simulation results are used to study this estimation. It was observed that the derived source power estimator will in general function well.Publication Data mining application: Customer retention at the Port of Singapore Authority (PSA)(1998-06) Ng, KianSing; Liu, Huan; Kwah, HweeBong; INFORMATION SYSTEMS & COMPUTER SCIENCE`Customer retention' is an important real-world problem in many sales and services related industries today. This work illustrates how we can integrate the various techniques of data-mining, such as decision-tree induction, deviation analysis and multiple concept-level association rules to form an intuitive and novel approach to gauging customer's loyalty and predicting their likelihood of defection. Immediate action taken against these `early-warnings' is often the key to the eventual retention or loss of the customers involved.Publication Neural-network feature selector(1997) Setiono, R.; Liu, H.; INFORMATION SYSTEMS & COMPUTER SCIENCEFeature selection is an integral part of most learning algorithms. Due to the existence of irrelevant and redundant attributes, by selecting only the relevant attributes of the data, higher predictive accuracy can be expected from a machine learning method. In this paper, we propose the use of a three-layer feedforward neural network to select those input attributes that are most useful for discriminating classes in a given set of input patterns. A network pruning algorithm is the foundation of the proposed algorithm. By adding a penalty term to the error function of the network, redundant network connections can be distinguished from those relevant ones by their small weights when the network training process has been completed. A simple criterion to remove an attribute based on the accuracy rate of the network is developed. The network is retrained after removal of an attribute, and the selection process is repeated until no attribute meets the criterion for removal. Our experimental results suggest that the proposed method works very well on a wide variety of classification problems. © 1997 IEEE.Publication On algorithms for digital signal processing of sequences(1996) Krishna Garg, H.; Ko, C.C.; Lin, K.Y.; Liu, H.; ELECTRICAL ENGINEERING; INFORMATION SYSTEMS & COMPUTER SCIENCEIn this work, we analyze the algebraic structure of fast algorithms for computing one- and two-dimensional convolutions of sequences defined over the fields of rational and complex rational numbers. The algorithms are based on factorization properties of polynomials and the direct sum property of modulo computation over such fields. Algorithms are described for cyclic as well as acyclic convolution. It is shown that under certain nonrestrictive conditions, all the previously defined algorithms over the fields of rational and complex rational numbers are also valid over the rings of finite integers. Examples are presented to illustrate the results.Publication Dimensionality reduction of unsupervised data(1997) Dash, M.; Liu, H.; Yao, J.; INFORMATION SYSTEMS & COMPUTER SCIENCEDimensionality reduction is an important problem for efficient handling of large databases. Many feature selection methods exist for supervised data having class information. Little work has been done for dimensionality reduction of unsupervised data in which class information is not available. Principal Component Analysis (PCA) is often used. However, PCA creates new features. It is difficult to obtain intuitive understanding of the data using the new features only. In this paper we are concerned with the problem of determining and choosing the important original features for unsupervised data. Our method is based on the observation that removing an irrelevant feature from the feature set may not change the underlying concept of the data, but not so otherwise. We propose an entropy measure for ranking features, and conduct extensive experiments to show that our method is able to find the important features. Also it compares well with a similar feature ranking method (Relief) that requires class information unlike our method.