Sutheera Puntheeranurak

Privacy-Preserving Collaborative Filtering Protocol Based on Similarity between Items

A recommendation system enables us to take information from huge datasets about tastes effectively. Many cryptographical protocols for computing privacy-preserving recommendation without leaking the privacy of users are proposed. However, the current issue is the large computational overhead depending the number of users. Hence, the application of the protocol is limited within small communities. In this paper, we address the issue of scalability by replacing the similarity between users by that of between items. Since the similarities between items can be publicly available, the recommendation steps are processed without dealing with confidential information such as the similarities between users. We propose an efficient scheme by using item-item similarities for providing a prediction of arbitrary values of rating. We show the performance and the accuracy evaluation of our proposed scheme based on a numerical experiment.

A new service matching definition and algorithm with SAWSDL

Semantic Web service (or briefly “SWS”) matching is a potential solution for automatic service discovery in various applications such as dynamic and automatic Web service composition. A number of approaches for SWS matching have been proposed. Most of them solely relied on concept subsumption relationship to improve the precision of Web service matching result from lexical-based methods. However such approaches suffer from several limitations such as not providing a fine-grained service matching degree, not supporting many-to-many matching between operation parameters, and not deriving data mappings between them. In this paper, we introduce a novel method for SWS matching based on SAWSDL, a semantic annotation specification for WSDL. We first define a fine-grained service matching result and then develop a corresponding service matching algorithm which can facilitate many-to-many matching of operation parameters as well as deriving concrete data mappings between them. Semantic techniques are used to infer and match data elements of operation parameters between a request and a service. The proposed algorithm was implemented and applied in the SeTEF — an automatic task-oriented business process execution system.

An Item-based collaborative filtering method using Item-based hybrid similarity

Item-based collaborative filtering is a preferred technique on recommender system. It uses a value of item rating similarity to predict users preference. In this paper, we include values of item attribute similarity to adjust the predicted rating equation for target item. The results of Item-based collaborative filtering that hybrid item rating similarity and item attribute similarity techniques have Mean Absolute Error (MAE) less than a traditional Item-based collaborative filtering technique and others. The proposed algorithm is efficient to predict better than traditional algorithm as shown in our experiments.

Hybrid Naive Bayes Classifier Weighting and Singular Value Decomposition Technique for Recommender System

Recommender System is tool for recommending products or services to customers which helps increase circulation products in electronic commerce systems. This paper proposes Naive Bayes Classifier Weighing Technique that applies to use with Singular Value Decomposition Technique for solving sparsity problem. The comparison Mean Absolute Error (MAE) between Hybrid Naive Bayes Classifier Weighing and Singular Value Decomposition Technique (HNBW-SVD) and Pure Singular Value Decomposition Technique (SVD) that found HNBW-SVD yield lower MAE than SVD.

SOA selection and replication based on QoS

Service Oriented Architecture (SOA) which contains a lot of consumers faces with reliability and availability problem. If one service fails, the user expects another alternative service to take the place of. In order to, increase service instance, service replication is adopted. To control the selection and replication of service, this paper proposes a service selection and replication based on Quality of Service (QoS). Service invocation will be even separated to each provider. While service replication will begin only if current service instances are not enough for number of consumers. Thus the system is able to control service replica properly and force service distribution for effecting regular consumer as few as possible.

A comparative study of relational database and key-value database for big data applications

Nowadays, Demands of web scale are in increasing and growing rapidly. Mobile applications, web technologies, social media always generates unstructured data that had lead to the advent of various NoSQL databases. Therefore, Big data applications are necessary to have an efficient technology to collect these data. However, a relational database is the traditional database that always uses in many applications and still has more valuable to play a significant role in the current information system. The main characteristics of NoSQL databases are schema-free, no relationship, no need to join as a relational database. The business organization expects that NoSQL database has better performance than a relational database. In This paper, we aim to compare the performance of Redis, which is a key-value database, one kind of NoSQL database, and MariaDB, which is a popular relational database. We designed a set of experiments with a large amount of data and compared the efficiency of the insert, update, delete and select transactions from various aspects on the same dataset. We measure the processing time of each transaction to evaluate the comparison. The results have shown that Redis has better runtime performance for insert, delete, update transaction under a specific condition or complex queries. MariaDB still is good for some conditions especially when we have a small data. Our study can help to choose a database that will be suitable for the real world applications because relational databases and NoSQL databases have different strengths and weakness.

Pathly: The integrated system for finding diverse journey methods

Public Transportation is a crucial factor for a journey not only the traveler but also people who live in Bangkok. From the variety of routes and many types of transport services that make a trip to Bangkok from one place to another would probably be shown by different ways. Pathly is our proposed algorithm which aims to develop the mobile applications that can suggest the diverse methods for a journey by the public transportation in Bangkok. It can make the appropriate choices considering from each condition and achieve user satisfaction. In our proposed, we implement the K-shortest path algorithm which is Yens algorithm for finding the diversity of routes. We design and collect all data from the graph database. We kept all data as a graph structure which represents the station as vertex and represents the path between 2 adjacent stations as an edge. Moreover, we make our consideration of 2 significant factors of a journey as time, fare over the trip. Pathly will help users to find the multiple methods to go to anywhere in Bangkok with an appropriate choice.

An improvement of video streaming service using dynamic routing over OpenFlow networks

Network technology is being developed as a next-generation network (NGN) which is a software-defined networking. The network management can be done directly from the control unit that separated from the transmission unit, called as a forwarder and decoupling controller, which makes efficient network management. In the near future, the next generation network will be used to replace a traditional network. This paper presents a framework of Quality of Service (QoS) for improvement video streaming service on Mesh Topology via OpenFlow framework. We used the optimal dynamic routing to distribute the flow of data. Our framework has enhanced quality for transferring video streaming when many clients request at the same time. In simulation results, which show that our proposed the QoS dynamic routing framework can achieve significant improvement in overall quality of streaming more than none QoS scenarios. In additional, the performance to maintain the server is good for clients even the system reaches the critical point of the network.

Hybrid recommender system with review helpfulness features

Nowadays, a recommendation system is an important technique in the development of electronic-commerce systems and the most popular approaches that use in many recommendation systems are a collaborative filtering algorithm. However, it still has problems such as scalability and sparse data. There are several previous methods used to deal with the weakness of collaborative filtering techniques such as a hybrid user model, but the results show their disadvantages in practical use. In this paper, we proposed a hybrid recommender system with review helpfulness features, which we have used to construct the hybrid model. In the experiment, the results of three recommendation techniques are compared: collaborative filtering based on hybrid user model, user-based collaborative filtering and our proposed method. The experimental results show that our proposed method is more efficient than other methods with the same dataset, in terms of accuracy. In addition, our proposed algorithm can decrease time consuming by constructing the hybrid model in the off-line phase and calculates the recommendation results in on-line phase.

An Enhanced User-Based Collaborative Filtering Recommendation System Using the Users’ Latent Relationships Weighting Utilization

Nowadays, A Recommendation system is an important technique in the development of electronic-commerce services and the most concerned approaches used in a recommendation system is a collaborative filtering algorithm, which uses the preference of users to make predictions. However, it works poorly to handle the sparse data. There are several previous methods used to deal with the weakness of collaborative filtering techniques such as the row-sampling approximating singular value decomposition algorithm, but the results show their disadvantages in practical use. In this paper, we propose an enhanced user-based collaborative filtering algorithm using users’ latent relationships weighting (CF-ULRW), which we have used in the predicted rating process. In the experiments, our proposed method is compared with the user-based collaborative filtering and the row-sampling approximating singular value decomposition. The experimental results show that our proposed method outperforms other methods with the same dataset.