Shie-Jue Lee

A multiple-kernel support vector regression approach for stock market price forecasting

Support vector regression has been applied to stock market forecasting problems. However, it is usually needed to tune manually the hyperparameters of the kernel functions. Multiple-kernel learning was developed to deal with this problem, by which the kernel matrix weights and Lagrange multipliers can be simultaneously derived through semidefinite programming. However, the amount of time and space required is very demanding. We develop a two-stage multiple-kernel learning algorithm by incorporating sequential minimal optimization and the gradient projection method. By this algorithm, advantages from different hyperparameter settings can be combined and overall system performance can be improved. Besides, the user need not specify the hyperparameter settings in advance, and trial-and-error for determining appropriate hyperparameter settings can then be avoided. Experimental results, obtained by running on datasets taken from Taiwan Capitalization Weighted Stock Index, show that our method performs better than other methods.

A Similarity Measure for Text Classification and Clustering

Measuring the similarity between documents is an important operation in the text processing field. In this paper, a new similarity measure is proposed. To compute the similarity between two documents with respect to a feature, the proposed measure takes the following three cases into account: a) The feature appears in both documents, b) the feature appears in only one document, and c) the feature appears in none of the documents. For the first case, the similarity increases as the difference between the two involved feature values decreases. Furthermore, the contribution of the difference is normally scaled. For the second case, a fixed value is contributed to the similarity. For the last case, the feature has no contribution to the similarity. The proposed measure is extended to gauge the similarity between two sets of documents. The effectiveness of our measure is evaluated on several real-world data sets for text classification and clustering problems. The results show that the performance obtained by the proposed measure is better than that achieved by other measures.

An Enhanced Type-Reduction Algorithm for Type-2 Fuzzy Sets

Karnik and Mendel proposed an algorithm to compute the centroid of an interval type-2 fuzzy set efficiently. Based on this algorithm, Liu developed a centroid type-reduction strategy to carry out type reduction for type-2 fuzzy sets. A type-2 fuzzy set is decomposed into a collection of interval type-2 fuzzy sets by -cuts. Then, the Karnik-Mendel algorithm is called for each interval type-2 fuzzy set iteratively. However, the initialization of the switch point in each application of the Karnik-Mendel algorithm is not a good one. In this paper, we present an improvement to Lius algorithm. We employ the previously obtained result to construct the starting values in the current application of the Karnik-Mendel algorithm. Convergence in each iteration, except the first one, can then speed up, and type reduction for type-2 fuzzy sets can be carried out faster. The efficiency of the improved algorithm is analyzed mathematically and demonstrated by experimental results.

A TSK-type neurofuzzy network approach to system modeling problems

We develop a neurofuzzy network technique to extract TSK-type fuzzy rules from a given set of input-output data for system modeling problems. Fuzzy clusters are generated incrementally from the training dataset, and similar clusters are merged dynamically together through input-similarity, output-similarity, and output-variance tests. The associated membership functions are defined with statistical means and deviations. Each cluster corresponds to a fuzzy IF-THEN rule, and the obtained rules can be further refined by a fuzzy neural network with a hybrid learning algorithm which combines a recursive singular value decomposition-based least squares estimator and the gradient descent method. The proposed technique has several advantages. The information about input and output data subspaces is considered simultaneously for cluster generation and merging. Membership functions match closely with and describe properly the real distribution of the training data points. Redundant clusters are combined, and the sensitivity to the input order of training data is reduced. Besides, generation of the whole set of clusters from the scratch can be avoided when new training data are considered.

FSKNN: Multi-label text categorization based on fuzzy similarity and k nearest neighbors

We propose an efficient approach, FSKNN, which employs fuzzy similarity measure (FSM) and k nearest neighbors (KNN), for multi-label text classification. One of the problems associated with KNN-like approaches is its demanding computational cost in finding the k nearest neighbors from all the training patterns. For FSKNN, FSM is used to group the training patterns into clusters. Then only the training documents in those clusters whose fuzzy similarities to the document exceed a predesignated threshold are considered in finding the k nearest neighbors for the document. An unseen document is labeled based on its k nearest neighbors using the maximum a posteriori estimate. Experimental results show that our proposed method can work more effectively than other methods.

Application of type-2 neuro-fuzzy modeling in stock price prediction

We present an application of type-2 neuro-fuzzy modeling to stock price prediction based on a given set of training data. Type-2 fuzzy rules can be generated automatically by a self-constructing clustering method and the obtained type-2 fuzzy rules cab be refined by a hybrid learning algorithm. The given training data set is partitioned into clusters through input-similarity and output-similarity tests, and a type-2 TSK rule is derived from each cluster to form a fuzzy rule base. Then the antecedent and consequent parameters associated with the rules are refined by particle swarm optimization and least squares estimation. Experimental results, obtained by running on several datasets taken from TAIEX and NASDAQ, demonstrate the effectiveness of the type-2 neuro-fuzzy modeling approach in stock price prediction.

Employing multiple-kernel support vector machines for counterfeit banknote recognition

Finding an efficient method to detect counterfeit banknotes is an imperative task in business transactions. In this paper, we propose a system based on multiple-kernel support vector machines for counterfeit banknote recognition. A support vector machine (SVM) to minimize false rates is developed. Each banknote is divided into partitions and the luminance histograms of the partitions are taken as the input of the system. Each partition is associated with its own kernels. Linearly weighted combination is adopted to combine multiple kernels into a combined matrix. Optimal weights with kernel matrices in the combination are obtained through semi-definite programming (SDP) learning. Two strategies are adopted to reduce the amount of time and space required by the SDP method. One strategy assumes the non-negativity of the kernel weights, and the other one is to set the sum of the weights to be unity. Experiments with Taiwanese banknotes show that the proposed approach outperforms single-kernel SVMs, standard SVMs with SDP, and multiple-SVM classifiers.

Deriving minimal conflict sets by CS-trees with mark set in diagnosis from first principles

To discriminate among all possible diagnoses using Hous theory of measurement in diagnosis from first principles, one has to derive all minimal conflict sets from a known conflict set. However, the result derived from Hous method depends on the order of node generation in CS-trees. We develop a derivation method with mark set to overcome this drawback of Hous method. We also show that our method is more efficient in the sense that no redundant tests have to be done. An enhancement to our method with the aid of extra information is presented. Finally, a discussion on top-down and bottom-up derivations is given.

A neuro-fuzzy approach for segmentation of human objects in image sequences

We propose a novel approach for segmentation of human objects, including face and body, in image sequences. Object segmentation is important for achieving a high compression ratio in modern video coding techniques, e.g., MPEG-4 and MPEG-7, and human objects are usually the main parts in the video streams of multimedia applications. Existing segmentation methods apply simple criteria to detect human objects, leading to the restriction of the usage or a high segmentation error. We combine temporal and spatial information and employ a neuro-fuzzy mechanism to overcome these difficulties. A fuzzy self-clustering technique is used to divide the base frame of a video stream into a set of segments which are then categorized as foreground or background based on a combination of multiple criteria. Then, human objects in the base frame and the remaining frames of the video stream are precisely located by a fuzzy neural network constructed with the fuzzy rules previously obtained and is trained by a singular value decomposition (SVD)-based hybrid learning algorithm. The proposed approach has been tested on several different video streams, and the results have shown that the approach can produce a much better segmentation than other methods.

Discovery of fuzzy temporal association rules

We propose a data mining system for discovering interesting temporal patterns from large databases. The mined patterns are expressed in fuzzy temporal association rules which satisfy the temporal requirements specified by the user. Temporal requirements specified by human beings tend to be ill-defined or uncertain. To deal with this kind of uncertainty, a fuzzy calendar algebra is developed to allow users to describe desired temporal requirements in fuzzy calendars easily and naturally. Fuzzy operations are provided and users can define complicated fuzzy calendars to discover the knowledge in the time intervals that are of interest to them. A border-based mining algorithm is proposed to find association rules incrementally. By keeping useful information of the database in a border, candidate itemsets can be computed in an efficient way. Updating of the discovered knowledge due to addition and deletion of transactions can also be done efficiently. The kept information can be used to help save the work of counting and unnecessary scans over the updated database can be avoided. Simulation results show the effectiveness of the proposed system. A performance comparison with other systems is also given.