Ladda Preechaveerakul

FIAST: A Novel Algorithm for Mining Frequent Itemsets

An efficient algorithm to mine frequent item sets is crucial for mining association rules. Most of the previously used algorithms have generally been developed for using the computational time effectively, reducing the number of candidate itemsets and decreasing the number of scan in the database. However, the time can be reduced by aggregate transactions having similar itemsets. This paper, then proposes an efficient algorithm for mining frequent item sets without generating candidate itemsets called FIAST (Frequent Itemsets Algorithm for Similar Transactions). The algorithm uses a divide-and-conquer method to reduce the task into the bitwise AND operation for finding itemsets and uses a depth-first search strategy to generate all frequent itemsets. The time complexity is analyzed by mathematical proof.

Information searching protocol: a smart protocol for specific content search over the Internet

Currently, information is very important to Internet users. Unfortunately, searching for specific information from the Internet is not easy as wishes. The existing search engine mechanisms cannot be performed using a pathname of URL as a search key. Therefore, users who have a partial pathname of URL cannot use their knowledge to narrow down the search results. Thus, users have to spend a long time searching for the required web site from the result list. This paper proposes a search protocol named Information Searching Protocol (ISP) that supports the multiple search contents for users who know a partial pathname of URL and keywords. Moreover, the architecture of the Global Search Engine System (GSES) that cooperates with the ISP and is responsible for the search mechanism is also proposed. The GSES consists of two separated parts: an Internet Search Protocol agent at the client site, and GSES components at the server site. These components allow users to perform the search using a pathname of URL composing with keywords. The functions of GSES components indicate that the ISP enhances the search mechanism. So, users receive more specific URL and can, shortly, get access to the required site.

A Novel Approach: Secure Information Notifying System Using RSS Technology

Internet affects our daily lives; people use it to contact each other. Many organizations also make use of it to supply information both general and sensitive information such as credit card information, financial business information to users as well as in using mobile devices. Really Simple Syndication (RSS) is also used to distribute the latest information over the Internet. However, the RSS technology does not have a mechanism to ensure that the incoming information is really secure. Therefore, we developed a Secure Information Notifying System with RSS Technology for Mobile Users (SInfoNS).

SIPE-SAP: SIP extension for signaling attacks protection

Voice over Internet Protocol (VoIP) has become popular as the Internet has been widely adopted. However, there are some new security issues faced by VoIP protocols. Using Session Initiation Protocol (SIP) as VoIP signaling protocol is introduced issues called signaling attacks. In this paper, we proposed a lightweight mechanism to protect SIP-based VoIP from signaling attacks by introducing an additional header field and a method to protect some important header fields such as Proxy-Authorization, To and Call-ID which are targeted to the attacks. The proposed mechanism protects against SIP messages forgery and modifying SIP messages. The experimental results show that the average registration response time of SIPE-SAP is not much different than normal SIP registration. In addition, the SIPE-SAP call setup overhead is less than TLS.

An image-based video copy detection using ordinal bitmap signature

This paper proposes a framework for video content copy detection for copyright or online video sharing. It utilizes a well-known ordinal measure and incorporates with a hierarchical bitmap indexing structure. This work improves on our previous work by introducing a few features: a new matching algorithm, black border removal, and selecting a suitable number of blocks. With these new features, the system has shown improvements by reducing the computational time and memory space in video matching process. The proposed system is able to handle several of video transformations including change in color, brightness, resolution, contrast, gamma, as well as the addition of noise, blurred, logo or text caption, flip (this case was failed to detect by our previous system). Moreover, by removing black borders before generating the signatures, the system yields better results in cases of ratio (letter/pillar box), zoom in and little crop/shift.

An Automatic Filtering Approach for News Video Feeds on TCP/IP-Based Mobile Devices

Really simple syndication (RSS) technology makes it for users possible to keep track of updated news. Podcasting is the technology based on RSS aims to aggregate multimedia files. For desktop users, viewing redundant RSS news video might not be a problem because of the high efficiency of the machine. But for mobile users, viewing redundant news video in a device with limited resource like PDAs or smart phones is wasting time, memory-space and battery-power. This paper proposes an automatic filtering for news video feeds for use on TCP/IP-based mobile devices. The system aggregates news video feeds from multiple sources and then filters an appropriate video to be shown on users mobile device. To reduce the processing on mobile devices, users are able to access news video through a Web-based reader.

DistEQ: Distributed equality query processing on Encoded Bitmap Index

With the next generation of data, called Big Data, an amount of data has grown rapidly, not only size but also variety and the speed of changing contents in its. The difficulties in Big Data are handling to store and process a large amount of data in real-time. Bitmap Indexes have widely known method to improve processing time. The advantage of these indexes is utilization of low-cost Boolean operations on the index before accessing to raw data. In this paper, we proposed a distributed and parallel processing, called DistEQ, to improve equality query on Encoded Bitmap Index by using MapReduce framework. Our comparative study demonstrates that our proposed approach outperforms other existing Encoded bitmap indexing techniques for an equality query in term of scalability and space-time trade-off.

Determining optimal retrieval points mechanism for RSS documents

Really Simple Syndication or RSS Technology helps users to receive updated contents using an aggregator from web sites. The aggregator is scheduled at time intervals to feed automatically. However, setting time intervals may cause a delay between the publication of new contents at a publisher site and the appearance at the aggregator, or no updated contents may occur. Then, we propose a mechanism to determine an optimal retrieval point for RSS feeds. The mechanism reduces the delay of retrieving contents. The study results show that the retrieval point can reduce 21% of the delay in retrieving contents in comparison with a previous research.

A novel Voting Algorithm of multi-class SVM for web page classification

The increasing numbers of web pages on the cyber world result to the less effectiveness of document retrieval that matches the need of users. The classification of web pages is one of the solutions to solve this problem. This paper proposes VAMSVM_WPC model which is a novel voting algorithm for classifying the web pages, which uses a multi-class SVM method. First, feature is generated from text and title, and then reduces the number of features by two feature selection techniques. Use these two types of features to give input to multi-class SVM. Finally, on the output of SVM, a voting algorithm is used to determine the category of the web pages. Results on CMU benchmark dataset show that using text and title feature with 1vsAll_Voting Algorithm gives the highest F-measure value.