Yih-Ling Hedley

Sampling, information extraction and summarisation of hidden web databases

Hidden Web databases maintain a collection of specialised documents, which are dynamically generated using page templates. This paper presents the Two-Phase Sampling (2PS) technique that detects and extracts query-related information from documents contained in databases. 2PS is based on a two-phase framework for the sampling, information extraction and summarisation of Hidden Web documents. In the first phase, 2PS samples and stores documents for further analysis. In the second phase, it detects Web page templates from sampled documents and extracts relevant information from which a content summary is then generated. Experimental results demonstrate that 2PS effectively eliminates irrelevant information from sampled documents and generates terms and frequencies with improved accuracy.

Query-related data extraction of hidden web documents

The larger amount of information on the Web is stored in document databases and is not indexed by general-purpose search engines (i.e., Google and Yahoo). Such information is dynamically generated through querying databases - which are referred to as Hidden Web databases. Documents returned in response to a user query are typically presented using template-generated Web pages. This paper proposes a novel approach that identifies Web page templates by analysing the textual contents and the adjacent tag structures of a document in order to extract query-related data. Preliminary results demonstrate that our approach effectively detects templates and retrieves data with high recall and precision.

A TNATS approach to hidden web documents

Hidden Web databases maintain a collection of documents, which are dynamically generated using Web page templates in response to user queries This paper presents a technique, Text with Neighbouring Adjacent Tag Segments (TNATS), to represent the contents of documents retrieved from an underlying database TNATS exploits tag structures that surround the textual content of a document This representation facilitates the process of detecting Web page templates and extraction of query-related information from documents We compare the performance of TNATS with existing techniques based on tag tree and text only representations Experimental results demonstrate that TNATS requires less processing time for information extraction than a tag tree representation It also produces optimum results in terms of detecting Web page templates and extracting query-related information.

Content summarisation of conversation in the context of virtual meetings: An enhanced TextRank approach

Organisations now frequently rely on virtual collaboration through the use of computer technology. After a sequence of meetings, participants may only need to refer to the most important points rather than the whole meeting proceedings. This paper addresses the need for automated meeting summarisation in virtual meeting systems. An extraction approach to summarisation is adopted and a new algorithm is proposed by extending the TextRank algorithm to include constructs representing the structure of the meeting. This helps extract the most relevant sentences from the meeting transcript. The proposed method was evaluated in the context of student-tutor meetings. Results show that harnessing and utilising the structure of a virtual meeting can lead to more relevant automated summaries.

Smartphone Based Human Activity and Postural Transition Classification with Deep Stacked Autoencoder Networks

Although traditionally used in the machine translation field, the encoder-decoder framework has been recently applied for the generation of video and image descriptions. The combination of Convolutional and Recurrent Neural Networks in these models has proven to outperform the previous state of the art, obtaining more accurate video descriptions. In this work we propose pushing further this model by introducing two contributions into the encoding stage. First, producing richer image representations by combining object and location information from Convolutional Neural Networks and second, introducing Bidirectional Recurrent Neural Networks for capturing both forward and backward temporal relationships in the input frames.