Md. Nasir Sulaiman

Development of Image Based Modeling for Determination of Oil Content and Days Estimation for Harvesting of Fresh Fruit Bunches

In this study, the relationship of oil extraction rate (OER) and fruit ripeness will be determined. The sample of oil palm fruits was collected from the unripe until the overripe stage and the oil content of the mesocarp for fresh fruit bunches (FFB) was extracted by using bunch analysis procedure to get the oil extraction rate. Using the same samples of FFB, the pixel value of images which measure in hue, was determined by developed image analysis software. The images were captured under an outdoor environment in an oil palm plantation. The sunlight intensity of environment was recorded using Extech light meter at various times of the day from morning to afternoon in the oil palm plantation. The result of the experiment that showed a good relationship was found between the oil content of FFB with its image pixel values. The mathematical model was developed in determining the optimum days for FFB harvesting.

K-Means Divide and Conquer Clustering

Cluster analysis, primitive exploration with little or no prior knowledge, consists of research developed across a wide variety of communities. Most clustering techniques ignore the fact about the different size or levels – where in most cases, clustering is more concern with grouping similar objects or samples together ignoring the fact that even though they are similar, they might be of different levels. For really large data sets, data reduction should be performed prior to applying the data-mining techniques which is usually performing dimension reduction, and the main question is whether some of these prepared and preprocessed data can be discarded without sacrificing the quality of results. Existing clustering techniques would normally merge small clusters with big ones, removing its identity. In this study we propose a method which uses divide and conquer technique to improve the performance of the K-Means clustering method.

A new clustering approach based on graph partitioning for navigation patterns mining

We present a study of the Web based user navigation patterns mining and propose a novel approach for clustering of user navigation patterns. The approach is based on the graph partitioning for modeling user navigation patterns. For the clustering of user navigation patterns we create an undirected graph based on connectivity between each pair of Web pages and we propose novel formula for assigning weights to edges in such a graph. The experimental results represent that the approach can improve the quality of clustering for user navigation pattern in Web usage mining systems. These results can be use for predicting userpsilas next request in the huge Web sites.

Adaptable Decentralized Service Oriented Architecture

In the Service Oriented Architecture (SOA), BPEL specified business processes are executed by non-scalable centralized orchestration engines. In order to address the scalability issue, decentralized orchestration engines are applied, which decentralize BPEL processes into static fragments at design time without considering runtime requirements. The fragments are then encapsulated into runtime components such as agents. There are a variety of attitudes towards workflow decentralization; however, only a few of them produce adaptable fragments with runtime environment. In this paper, producing runtime adaptable fragments is presented in two aspects. The first one is frequent-path adaptability that is equal to finding closely interrelated activities and encapsulating them in the same fragment to omit the communication cost of the activities. Another aspect is proportional-fragment adaptability, which is analogous to the proportionality of produced fragments with number of workflow engine machines. It extenuates the internal communication among the fragments on the same machine. An ever-changing runtime environment along with the mentioned adaptability aspects may result in producing a variety of process versions at runtime. Thus, an Adaptable and Decentralized Workflow Execution Framework (ADWEF) is introduced that proposes an abstraction of adaptable decentralization in the SOA orchestration layer. Furthermore, ADWEF architectures Type-1 and Type-2 are presented to support the execution of fragments created by two decentralization methods, which produce customized fragments known as Hierarchical Process Decentralization (HPD) and Hierarchical Intelligent Process Decentralization (HIPD). However, mapping the current system conditions to a suitable decentralization method is considered as future work. Evaluations of the ADWEF decentralization methods substantiate both adaptability aspects and demonstrate a range of improvements in response-time, throughput, and bandwidth-usage compared to previous methods.

Modeling of oil palm fruit maturity for the development of an outdoor vision system.

Color is the most important indicator farmers use to determine the maturity of the oil palm fruit called fresh fruit bunches (FFB) in the manual harvesting process. To automate the harvesting operation, the development of a vision system will replace the human eye for mature FFB recognition. In real plantation environments, variations in the daylight caused the light intensity to change, thus becoming the main issue that affects the automatic recognition process. In this study, the matured FFB was captured using a Sony digital Handycam on the day shift period. At the same time period of daylight intensity, a unit on foot candles (FC) also was simultaneously recorded using an Extech lightmeter data logger. From the linear regression analysis process, the mathematical model shows that there is a linear change between daylight intensity with the pixel value of the components green and blue. For the pixel value of the red component, the value will be linear at a maximum of 255 and at a certain intensity. To validate the mathematical model, this equation is used in the development of software for outdoor recognition processes.

A new classification model for online predicting users’ future movements

Nowadays many internet users prefer to navigate their interest web pages in special web site rather than navigating all web pages in the web site. For this reason some techniques have been developed for predicting user’s future requests. Data manning algorithms can be applied to many prediction problems. We can exploit Web Usage Mining for Knowledge extracting based on user behavior during the web navigation. The WUM applies data mining techniques for extracting knowledge from user log files in the particular web server. The WUM can model user behavior and, therefore, to forecast their future movements by mining user navigation patterns. To provide online prediction efficiently, we advance architecture for online predicting in web usage mining system by proposing novel model based on Longest Common Subsequence algorithm for classifying user navigation patterns. The prediction of users’ future movements by this manner can improve accuracy of recommendations.

Fast Discovery Of Long Patterns For Association Rules

The most time consuming process in discovering association rules is identifying the frequent patterns especially in the cases when the database contains long patterns. An algorithm called Flex for identifying frequent patterns especially efficient when the patterns are long is proposed by successive construction of the nodes lexicographic tree. The vertical counting strategy to facilitate fast discovery is used in support computation. The experimental result shows that Flex outperform Apriori, a well-known and widely used algorithm for patterns discovery.

Integrative Gene Selection for Classification of Microarray Data

Microarray data classification is one of the major interests in health informatics that aims at discovering hidden patterns in gene expression profiles. The main challenge in building this classification system is the curse of dimensionality problem. Thus, there is a considerable amount of studies on gene selection method for building effective classification models. However, most of the approaches consider solely on gene expression values, and as a result, the selected genes might not be biologically meaningful. This paper presents an integrative gene selection for improving microarray data classification performance. The proposed approach employs the association analysis technique to integrate both gene expression and biological data in identifying informative genes. The experimental results show that the proposed gene selection outperformed the traditional method in terms of accuracy and number of selected genes.

Propositional satisfiability algorithm to find minimal reducts for data mining

A fundamental problem in data mining is whether the whole information available is always necessary to represent the information system (IS). Reduct is a rough set approach in data mining that determines the set of important attributes to represent the IS. The search for minimal reduct is based on the assumption that within the dataset in an IS, there are attributes that are more important than the rest. An algorithm in finding minimal reducts based on Propositional Satisfiability (SAT) algorithm is proposed. A branch and bound algorithm is presented to solve the proposed SAT problem. The experimental result shows that the proposed algorithm has significantly reduced the number of rules generated from the obtained reducts with high percentage of classification accuracy.

Decentralized replication strategies for P2P based Scientific Data Grid

Scientific Data Grid provides geographically distributed resources for large-scale data-intensive applications that generate large scientific data sets and it mostly deals with large computational problems. Research in the area of grid has given various ideas and solutions to address these requirements. However, since the number of participants (scientists and institutes) that involve in this kind of environment is increasing tremendously, scalability, availability and reliability have been the core problem for such system. Peer-to-peer (P2P) is one of the architecture that promising scale and dynamism environment. In this paper, we present a P2P model for Scientific Data Grid that utilizes the P2P services to address those problems. For the purpose of this study, we have developed and used our own data grid simulation written using PARSEC. In this paper, we illustrate our P2P Scientific Data Grid model, our data grid simulation and the design of proposed data replication strategies. We then analyze the performance of data discovery service with and without the existence of replication strategies relative to their success rates, response time, average number of hop and bandwidth consumption. The results from simulation study that show how the proposed replication strategies promote high data availability in the proposed Scientific Data Grid model and how these strategies improve the discovery process are presented.