Razali Yaakob

An investigation of an evolutionary approach to the opening of Go

The game of Go can be divided into three stages; the opening, the middle, and the end game. In this paper, evolutionary neural networks, evolved via an evolutionary strategy, are used to develop opening game playing strategies for the game. Go is typically played on one of three different board sizes, i.e., 9/spl times/9, 13/spl times/13 and 19/spl times/19. A 19/spl times/19 board is the standard size for tournament play but 9/spl times/9 and 13/spl times/13 boards are usually used by less-experienced players or for faster games. This work focuses on the opening, using a 13/spl times/13 board. A feed forward neural network player is played against a static player (Gondo), for the first 30 moves. Then Gondo takes the part of both players to play out the remainder of the game. Two experiments are presented which indicate that learning is taking place.

An extended ID3 decision tree algorithm for spatial data

Utilizing data mining tasks such as classification on spatial data is more complex than those on non-spatial data. It is because spatial data mining algorithms have to consider not only objects of interest itself but also neighbours of the objects in order to extract useful and interesting patterns. One of classification algorithms namely the ID3 algorithm which originally designed for a non-spatial dataset has been improved by other researchers in the previous work to construct a spatial decision tree from a spatial dataset containing polygon features only. The objective of this paper is to propose a new spatial decision tree algorithm based on the ID3 algorithm for discrete features represented in points, lines and polygons. As in the ID3 algorithm that use information gain in the attribute selection, the proposed algorithm uses the spatial information gain to choose the best splitting layer from a set of explanatory layers. The new formula for spatial information gain is proposed using spatial measures for point, line and polygon features. Empirical result demonstrates that the proposed algorithm can be used to join two spatial objects in constructing spatial decision trees on small spatial dataset. The proposed algorithm has been applied to the real spatial dataset consisting of point and polygon features. The result is a spatial decision tree with 138 leaves and the accuracy is 74.72%.

Classification model for hotspot occurrences using spatial decision tree algorithm

Developing a predictive model for forest fires occurrence is an important activity in a fire prevention program. The model describes characteristics of areas where fires occur based on past fires data. It is essential as an early warning system for preventing forest fires, thus major damages because of fires can be avoided. This study describes the application of data mining technique namely decision tree on forest fires data. We improved the ID3 decision tree algorithm such that it can be utilized on spatial data in order to develop a classification model for hotspots occurrence. The ID3 algorithm which is originally designed for a non-spatial dataset has been improved to construct a spatial decision tree from a spatial dataset containing discrete features (points, lines and polygons). As the ID3 algorithm that uses information gain in the attribute selection, the proposed algorithm uses spatial information gain to choose the best splitting layer from a set of explanatory layers. The new formula for spatial information gain is proposed using spatial measures for point, line and polygon features. The proposed algorithm has been applied on the forest fire dataset for Rokan Hilir district in Riau Province in Indonesia. The dataset contains physical data, socio-economic, weather data as well as hotspots and non-hotspots occurrence as target objects. The result is a spatial decision tree with 276 leaves with distance from target objects to the nearest river as the first test layer and the accuracy on the training set of 87.69%. Empirical result demonstrates that the proposed algorithm can be used to join two spatial objects in constructing a spatial decision tree from a spatial dataset. The algorithm results a predictive model for hotspots occurrence from the real dataset on forest fires with high accuracy on the training set.

A framework for evaluating skyline query over uncertain autonomous databases

The perception of skyline query is to find a set of objects that is much preferred in all dimensions. While this theory is easily applicable on certain and complete database, however, when it comes to data integration of databases where each has different representation of data in a same dimension, it would be difficult to determine the dominance relation between the underlying data. In this paper, we propose a framework, SkyQUD, to efficiently compute the skyline probability of datasets in uncertain dimensions. We explore the effects of having datasets with uncertain dimensions in relation to the dominance relation theory and propose a framework that is able to support skyline queries on this type of datasets.

A hybrid model using genetic algorithm and neural network for predicting dengue outbreak

Prediction of dengue outbreak becomes crucial in Malaysia because this infectious disease remains one of the main health issues in the country. Malaysia has a good surveillance system but there have been insufficient findings on suitable model to predict future outbreaks. While there are previous studies on dengue prediction models in Malaysia, unfortunately some of these models still have constraints in finding good parameter with high accuracy. The aim of this paper is to design a more promising model for predicting dengue outbreak by using a hybrid model based on genetic algorithm for the determination of weight in neural network model. Several model architectures are designed and the parameters are adjusted to achieve optimal prediction performance. Sample data that covers dengue and rainfall data of five districts in Selangor collected from State Health Department of Selangor (SHD) and Malaysian Meteorological Department is used as a case study to evaluate the proposed model. However, due to incomplete collection of real data, a sample data with similar behavior was created for the purpose of preliminary experiment. The result shows that the hybrid model produces the better prediction compared to standalone models.

Multi-View Human Action Recognition Using Wavelet Data Reduction and Multi-Class Classification☆

Human action recognition from video has several potential to apply in different real-life applications, but the most cases in this field suffer from the variation in viewpoint. Most of published methods in this area are considered the performance of each single camera, therefore the change in the viewpoints significantly decrease the recognition rate. In this paper, multiple views are considered together and a method has proposed to recognize human action depicted in multi-view image sequences. In the first step, the border of the human bodys silhouette is extracted and distance signal is calculated. In the next step, the wavelet transform is applied to extract coefficients of single-view features, and then the extracted features are combined to compose multi-view features. Finally a hierarchical classifier using support vector machine and Naive Bayes classifiers is implemented to classify the actions. The average of overall action recognition accuracy for 12 actions using 5 different angles of views on the IXMAS dataset is 88.22. The results of experiments on the popular multi-view dataset have shown the proposed method achieves high and state-of-the-art success rates. In other word, combination of single-view extracted features from the wavelet approximation coefficients and composing the multi-view features can be used as the multi-view features. Further, the hierarchical classifier can be applied to recognize actions in multi-view human action recognition area.

Silhouette-based multi-view human action recognition in video

In this paper, a human action recognition method is presented where pose features are represented using contour points of the human silhouette, and actions are learned by using sequences of multi-view contour points. The differences and divergences among actors performing the same action are handled by considering variations in shape and speed. Experimental results on the IXMAS dataset show promising success rates, exceeding that of existing multi-view human action recognition state-of-the-art techniques.

Blog quality model

Breakthroughs in technology are making the internet an ever-growing phenomenon, and we have witnessed an enormous growth of blogs in the blogosphere. However, the blogosphere has been crippled by disorganised and uncontrolled growth, and many blogs are of poor quality. Development domains, such as software engineering, website engineering, and information systems, have provided accepted models for the assessment of the quality of their products. However, to the best of our knowledge, there appears to be no standard model for the assessment of blog quality. In this paper, we propose a blog quality model as a guide for bloggers at large, with a set of 49 criteria grouped into 11 families of features that are relevant to blog quality. This model has been constructed by determining a set of criteria from a review of the relevant literature and blogs and then measuring the acceptability of these criteria by means of questionnaire surveys sent to sample populations of blog readers.

Application of Rasch Model in Validating the Content of Measurement Instrument for Blog Quality

Research in blog quality is very crucial nowadays in order to have a good quality blog in the blogosphere. The blog quality criteria have been derived from a rigorous metadata analysis. Yet, these criteria have not been reviewed and their significance has not been proven systematically. In this paper, Rasch Model is applied to produce an empirical evidence of content validity of the blog quality criteria. This study confirms that the definitions of 11 families and the 49 criteria assigned have content validity by mean of online survey. These criteria will then be used as a basis of constructing the instrument to measure the acceptability of the criteria for blog quality.

Modeling forest fires risk using spatial decision tree

Forest fires have long been annual events in many parts of Sumatra Indonesia during the dry season. Riau Province is one of the regions in Sumatra where forest fires seriously occur every year mostly because of human factors both on purposes and accidently. Forest fire models have been developed for certain area using the weightage and criterion of variables that involve the subjective and qualitative judging for variables. Determining the weights for each criterion is based on expert knowledge or the previous experienced of the developers that may result too subjective models. In addition, criteria evaluation and weighting method are most applied to evaluate the small problem containing few criteria. This paper presents our initial work in developing a spatial decision tree using the spatial ID3 algorithm and Spatial Join Index applied in the SCART (Spatial Classification and Regression Trees) algorithm. The algorithm is applied on historic forest fires data for a district in Riau namely Rokan Hilir to develop a model for forest fires risk. The modeling forest fire risk includes variables related to physical as well as social and economic. The result is a spatial decision tree containing 138 leaves with distance to nearest river as the first test attribute.