Section: Overall Objectives
Keywords : information system, user-centered design, evaluation, knowledge discovery, KDD, data mining, usage mining, document mining, knowledge management, ontology management, information retrieval, recommender system, Web mining, semantic Data mining, Semantic Web, personnalisation data stream mining, transportation system, World Wide Web, Security, intrusion detection, anomaly detection.
Overall Objectives
AxIS is carrying out research in the area of Information and Knowledge Systems (ISs) with a special interest in evolving large ISs such as Web based-information Systems. Our ultimate goal is to improve the overall quality of ISs, to support designers during the design process and to ensure ease of use to end users. We are convinced that to reach this goal, according to the constant evolution of actual and future ISs, it is necessary to anticipate the usage and its analysis and also the maintenance very early in the design process.
To achieve such a research, we have set up in July 2003 a multidisciplinary team that involves people from different computer sciences domains (Artificial Intelligence, Data Mining & Analysis, Software Engineering, Document Management from 2004) and at the end of 2005 from Ergonomics (Human Sciences), all of them focusing on information systems. Our goal is of course to improve efficiency of machine learning and data mining methods but also to improve the quality of results. The originality of AxIS project-team is to adopt a cognitive and inter-disciplinary approach for the whole KDD process and for each step (preprocessing, data mining, interpretation).
To address this challenge, relying on our scientific foundations (see our 2007 activity report,section 3 or http://ralyx.inria.fr/2007/Raweb/axis/uid0.html ), we had a first 4 years steps dedicated to the design of methodological and technical building blocks for IS mining (usage, content and structure). Our next steps started this year identified three applicative objectives:
-
Analysing Collective Usage: the need to understand the usage of users or a group of users becomes more and more important in the context of organisations (academic or not) or enterprises for example. Building new KDD tools, able to address the real complexity of mining, that is mining from various sources which become larger and larger seems crucial to us. So is the capacity to adopt pluridisciplinary approaches.
-
Offering new kinds of support tools for Information/knowledge retrieval and management: to face the increasing amount of available information/knowledge, tools for retrieving relevant documents have been built. In addition to theses tools, we believe that sophisticated tools are needed for supporting all kinds of users or groups of users in discovering new information/knowledge and more generally in managing their own information/knowledge. These new tools will have to be very flexible KDD based management tools a) for the end-users, they will support the annotation of their selected documents, exhibit and synthesize new information from large documents and selected data collections (clusters of documents for example) or b) for others types of actors such as webmasters or any responsible, they will help them in evaluating and maintaining the IS or in analysing the usage of the IS or a group of users, etc.
-
Supporting the evolution of the future ISs (or a collection of XML documents) and of their use: future information systems will evolve faster and faster and will become larger and larger. To support the validation/evaluation of evolving information systems (in terms of content and structure), we believe in the design of powerful maintenance tools based on formal methods from software engineering. Such an evolution might also be due to the evolution of the usage by the end-users. Detecting such behavioural changes will be very useful in order to offer information retrieval support tools able to take into account small usage practice changes. They will be adaptable to the user profile/session and adaptive (i.e with some machine learning capabilities).
Our researches are organised in three topics according to the genericity level of results:
