Tran Duc Minh Email & Phone Number

Sophia Antipolish, France

Deepspa Project

Nice Area, France

Subject: Discovering multi-relational association rules from ontological knowledge bases to enrich ontologies.In this project, based on ideas derived from ILP (Inductive Logic Programming), the project aim at discovering hidden knowledge patterns in the form of multi-relational association rules by exploiting the evidence coming from the assertional data of ontological knowledge bases. Specifically, discovered rules are coded in SWRL (Semantic Web Rule Language) to easily integrated within the ontology, thus enriching its expressive power (TBox) and augmenting the assertional knowledge (ABox) that can be derived.Achievements:Two algorithms were created to apply to ontological knowledge bases which are proposed for finding rules with a high inductive power:1. Level-wise generated-and-test algorithm 2. Evolutionary algorithm.In addition, In order to improve the system performance, an algorithm is proposed to compute metrics used to evaluate the quality of the rules instead of querying via SPARQL-DL. This algorithm based on the hash join algorithm used in the implementation of RDBMS (Relational Database Management System) but is adapted to fit the notions in the semantic web.Experiments were performed on publicly available ontologies related to Financial, Biological, ... , validating the performances of approach and comparing them with previous systems.The content of the project is published in international conferences: EvoStar, GECCO, EKAW, SACTechnical environment :Languages: JavaOS: UbuntuData: XML Database, RDF, OWLTools: Subversion, Perlet, HermiTDesign: ProtégeParallel System: Hadoop

Hanoi, Vietnam

Teaching courses related to C++, Java, Data Structure and Artificial Intelligent.Achievements:Give lectures and supervise 4th-5th year students in some graduate projects related to semantic webs, artificial intelligent, database management.Technical environment:Languages: C++, Java Tools: Artificial Intelligent, Data Structure

Hanoi - Vietnam

This project aims at developing a system ERM (Electronic Records Management) for government agencies to shorten records processing time. submit their applications to the government agency. Records will be transferred to some departments for processing by a local network. Leaders can approve records by desktop along with electronic signature. Citizens can track their dossier processing steps on the Internet.The approach of the project is as follows:1. The system uses a centralized database (MS-SQL) and all computers are connected to this database to use the service.2. Profile analysis and design of the project uses a visual modelling language UML.3. Applications for leadership, records department and citizens are written in C# using MVC architecture and GUI is web-based.4. Software is estimated by Use Case Points method to forecast the software size.Achievements:The project has deployed in some provincial agencies in Vietnam and has achieved the desired results.Technical environment:Languages: C#OS: WindowsDatabase: MS-SQLTools: UMLMethodology: MVCSoftware estimation: UCP

Hanoi - Vietnam

The goal of the project is provide an application that can predict the trend of the stock market. Data is analyzed based on candlestick patterns and trading volume. The approach of the project is as follows:1. Build a simple web crawler in Java along with Selenium WebDriver to retrieve the whole stock data for 15 years. Retrieved data is transformed into candlestick patterns and stored in an MS-SQL database.2. Build an algorithm based on genetic algorithm to compute, statistics and forecast the market trend in the future. The best results of the algorithm will be stored back to MS-SQL database to support future prediction. The algorithm is all written in C++ language because algorithm support libraries are written in C++, the data is extracted from MS-SQL database via ODBC connection. Achievements:This project is a new approach that uses artificial intelligence to analyze candlestick patterns, trading volume and give the trend of the market. The best predicted outputs range from 70-85% The results of the project are used to support the research of students in both faculties (Information Technology and Finance) to quickly access knowledge in two areas of Arrificial Intelligent and Market Analysis.Technical environment :Languages: C++/JavaOS: Ubuntu, WindowsDatabase: MS-SQLTools: Selenium WebDriver

Hanoi - Vietnam

This project is a module in an overall project related to student management system. This module needs to identify 10 different speech signals (numbers from 0 to 9) to support the telephone switchboard. Users use speech (digits) to interact with the system through the phone. The approach of the project is as follows:1. Audio data is digitized by an open source software written in C++.2. Build an algorithm based on fuzzy logic to identify speech. The algorithm is all written in Visual C++. Achievements:An algorithm was created to identify the numbers from 0 to 9 in Vietnamese to support the telephone switchboard. Algorithms are packaged in a library (.dll) so that other members in the overall project can use it.The module has been deployed in practice with up to 95% accuracy. Module Demo: https://www.youtube.com/watch?v=8LMyJaGpyRg Technical environment:Languages: C++ OS: WindowsTools: MFCMethodology: Fuzzy Logic

Doctor Of Philosophy - Phd, Computer Science

Subject of thesis: Discovering multi-relational association rules from ontological knowledge bases to enrich ontologies.

Master'S Degree, Computer Science

Bachelor'S Degree, Computer Science