Tracking and Measurement for Search Quality

I built a job search engine using Elasticsearch to illustrate the role of tracking, measurement and evaluation in search. I described the process in Search Engineering 101. To download the source, proceed to Github.

Language Analysis Essentials (LANES)

Over the years, I have work on a variety of NLP and text mining techniques for extracting and organising knowledge from unstructured text. I have identified a number of essential ones and released their source codes under the GNU GPL v3 licence. To download, proceed to Github.

Conversational System using Question-Answer Pairs mined from the Web

The system uses innovative dialogue strategies and question-answer pairs mined from social sites on the Web to carry out coherent conversations with humans on broad range of topics.

Real-Time Structuring of Text using Ontologies for Intelligent Text Editing

A tool for editing text and exploring related documents across different domains using automatically constructed lightweight ontologies. Refer to my PhD dissertation for more information.

Information Exploration using Automatically Constructed Ontologies

The interface use lightweight ontologies, which are automatically constructed from unstructured webpages, to organise documents for search and exploration. Refer to my PhD dissertation for more information.

Real-Time Clinical Progress Note Cleaning and Structuring

A tool for cleaning and normalising progress notes on-the-fly as clinicians compose them. The tool uses a combination of innovative spelling error correction, abbreviation expansion and other natural language processing techniques from my research.

Real-Time Prescription Text Validation

A tool for reducing two common prescribing errors, namely, incorrect dosage/frequency and drug name confusion. The tool performs on-the-fly prescription bracketing to structure free-text prescriptions into different components, and prescribing habit analysis to analyse the structured prescriptions and alerts the physicians of any potential medication errors. Refer to my AIM Journal article.

Detection of Upper Bodies from Image Sequences

The tool detects upper bodies from unconstrained still images. It combines the reliability of face detectors with the robustness of people detection based on HoG descriptors to improve the accuracy of upper body detection from monocular still images with cluttered background, poor illumination, motion blur and high-degree of occlusion. Refer to my ICIEA paper for details