Yuhanis Yusof

Enhanced artificial bee colony for training least squares support vector machines in commodity price forecasting

Abstract The importance of optimizing machine learning control parameters has motivated researchers to investigate for proficient optimization techniques. In this study, a Swarm Intelligence approach, namely artificial bee colony (ABC) is utilized to optimize parameters of least squares support vector machines. Considering critical issues such as enriching the searching strategy and preventing over fitting, two modifications to the original ABC are introduced. By using commodities prices time series as empirical data, the proposed technique is compared against two techniques, including Back Propagation Neural Network and by Genetic Algorithm. Empirical results show the capability of the proposed technique in producing higher prediction accuracy for the prices of interested time series data.

Dengue Outbreak Prediction: A Least Squares Support Vector Machines Approach

Dengue fever (DF) and the potentially fatal dengue hemorrhagic fever (DHF) continue to be a crucial public health concern in Malaysia. This paper proposes a prediction model that incorporates Least Squares Support Vector Machines (LS-SVM) in predicting future dengue outbreak. Data sets used in the undertaken study includes data on dengue cases and rainfall level collected in five districts in Selangor. Data were preprocessed using the Decimal Point Normalization before being fed into the training model. Prediction results of unseen data show that the LS-SVM prediction model outperformed the Neural Network model in terms of prediction accuracy and computational time.

Weight-Based Firefly Algorithm for Document Clustering

Existing clustering techniques have many drawbacks and this includes being trapped in a local optima. In this paper, we introduce the utilization of a new meta-heuristics algorithm, namely the Firefly algorithm (FA) to increase solution diversity. FA is a nature-inspired algorithm that is used in many optimization problems. The FA is realized in document clustering by executing it on Reuters-21578 database. The algorithm identifies documents that has the highest light intensity in a search space and represents it as a centroid. This is followed by recognizing similar documents using the cosine similarity function. Documents that are similar to the centroid are located into one cluster and dissimilar in the other. Experiments performed on the chosen dataset produce high values of Purity and F-measure. Hence, suggesting that the proposed Firefly algorithm is a possible approach in document clustering.

Forecasting Model Based on LSSVM and ABC for Natural Resource Commodity

 Abstract—Reliable forecasts of the price of natural resource commodity is of interest for a wide range of applications. This includes generating macroeconomic projections and in assessing macroeconomic risks. Various approaches have been introduced in developing the required forecasting models. In this paper, a forecasting model based on an optimized Least Squares Support Vector Machine is proposed. The determination of hyper-parameters is performed using a nature inspired algorithm, Artificial Bee Colony. The proposed forecasting model is realized in gold price forecasting. The undertaken experiments showed that the prediction accuracy and Mean Absolute Percentage Error produced by the proposed model is better compared to the one produced using Least Squares Support Vector Machine as an individual.

Document Clustering Based on Firefly Algorithm

Document clustering is widely used in Information Retrieval however, existing clustering techniques suffer from local optima problem in determining the k number of clusters. Various efforts have been put to address such drawback and this includes the utilization of swarm-based algorithms such as particle swarm optimization and Ant Colony Optimization. This study explores the adaptation of another swarm algorithm which is the Firefly Algorithm (FA) in text clustering. We present two variants of FA; Weight- based Firefly Algorithm (WFA) and Weight-based Firefly Algorithm II (WFAII). The difference between the two algorithms is that the WFAII, includes a more restricted condition in determining members of a cluster. The proposed FA methods are later evaluated using the 20Newsgroups dataset. Experimental results on the quality of clustering between the two FA variants are presented and are later compared against the one produced by particle swarm optimization, K-means and the hybrid of FA and -K-means. The obtained results demonstrated that the WFAII outperformed the WFA, PSO, K-means and FA-Kmeans. This result indicates that a better clustering can be obtained once the exploitation of a search solution is improved.

Application of LSSVM by ABC in energy commodity price forecasting

The importance of the hyper parameters selection for a kernel-based algorithm, viz. Least Squares Support Vector Machines (LSSVM) has been a critical concern in literature. In order to meet the requirement, this work utilizes a variant of Artificial Bee Colony (known as mABC) for hyper parameters selection of LSSVM. The mABC contributes in the exploitation process of the artificial bees and is based on Levy mutation. Realized in crude oil price forecasting, the performance of mABC-LSSVM is guided based on Mean Absolute Percentage Error (MAPE) and Root Mean Square Error (RMSPE) and compared against the standard ABC-LSSVM and LSSVM optimized by Genetic Algorithm. Empirical results suggested that the mABC-LSSVM is superior than the chosen benchmark algorithms.

A Newton’s Universal Gravitation Inspired Firefly Algorithm for Document Clustering

The divisive clustering has the advantage to build a hierarchical structure that is more efficient to represent documents in search engines. Its operation employs one of the partition clustering algorithms that leads to being trapped in a local optima. This paper proposes a Firefly algorithm that is based on Newton’s law of universal gravitation, known as Gravitation Firefly Algorithm (GFA), for document clustering. GFA is used to find centers of clusters based on objective function that maximizes the force between each document and an initial center. Upon identification of a center, the algorithm then locates documents that are similar to the center using cosine similarity function. The process of finding centers for new clusters continues by sorting the light intensity values of the balance documents. Experimental results on Reuters datasets showed that the proposed Newton inspired Firefly algorithm is suitable to be used for document clustering in text mining.

Short Term Traffic Forecasting Based on Hybrid of Firefly Algorithm and Least Squares Support Vector Machine

The goal of an active traffic management is to manage congestion based on current and predicted traffic conditions. This can be achieved by utilizing traffic historical data to forecast the traffic flow which later supports travellers for a better journey planning. In this study, a new method that integrates Firefly algorithm (FA) with Least Squares Support Vector Machine (LSSVM) is proposed for short term traffic speed forecasting, which is later termed as FA-LSSVM. In particular, the Firefly algorithm which has the advantage in global search is used to optimize the hyper-parameters of LSSVM for efficient data training. Experimental result indicates that the proposed FA-LSSVM generates lower error rate and a higher accuracy compared to a non-optimized LSSVM. Such a scenario indicates that FA-LSSVM would be a competitor method in the area of time series forecasting.

Replication in data grid: Determining important resources

Replication is an important activity in determining the availability of resources in data grid. Nevertheless, due to high computational and storage cost, having replicas for all existing resources may not be an efficient practice. Existing approach in data replication have been focusing on utilizing information on the resource itself or network capability in order to determine replication of resources. In this paper, we present the integration of three types of relationships for the mentioned purpose. The undertaken approach combines the viewpoint of user, file system and the grid itself in identifying important resource that requires replication. Experimental work has been done via OptorSim and evaluation is made based on the job execution time. Results suggested that the proposed strategy produces a better outcome compared to existing approaches.

PARTIAL RULE MATCH FOR FILTERING RULES IN ASSOCIATIVE CLASSIFICATION

In this study, we propose a new method to enhance the accuracy of Modified Multi-class Classification based on Association Rule (MMCAR) classifier. We introduce a Partial Rule Match Filtering (PRMF) method that allows a minimal match of the items in the rule’s body in order for the rule to be added into a classifier. Experiments on Reuters-21578 data sets are performed in order to evaluate the effectiveness of PRMF in MMCAR. Results show that the MMCAR classifier performs better as compared to the chosen competitors.