The work plan of the FP 7 Support Action NHECD is strucutred in 4 technical work packages (WPs):
WP1's objectives for the first year are to design and develop a solid self-maintained framework able to sustain a large and growing repository of scientific research papers. The framework will be based on commercial best of breed tools for enterprise content management. On top of these tools, WP1 activities will concentrate on the design and development of a customized software for automated enrichment of the toxicology taxonomy and for automated extraction of results from the scientific publications. As of the second year, a stable operation is expected that will enable to build up a very large knowledge base of scientific publications. Subsequently, the operations mode will focus on random reviews of the extracted results, making corrective actions per need along with scheduled enhancements to the software. Three software service packs, including enhancements and updates, after the first two years of the project are envisaged.
The toxicology taxonomy will be reviewed and corrected by domain experts from TAU, JRC and IVAM in WP2. At first, a prototype for the data base taxonomy will be set up. The taxonomic structure will be derived almost entirely from data contained already in a data base, according to the relations cast by its data modeling. Taking a cue from data mining, a data structure will be employed to capture data in order to produce nested topic lists for the taxonomy. The combination of this data structure with its nested mapping algorithms is able to transform selected lists of data, or information, into this new type of knowledge representation. End-users shall view this data base taxonomy from an interface that functions like data bases' navigation structure. These advances redefine the current data base retrieval capabilities, by enabling users to browse and explore data base content by visualizing the data and its relations along paths supported by a conceptual model. More importantly, these advances enable end-users to navigate these access paths, and to pinpoint and retrieve with high level of accuracy the information needed.
The extracted results, obtained from the automated taxonomy (WP1), will be ranked and commented according to parameters defined by domain experts from TAU, JRC and IVAM. The ranking of the loaded meta-data will be based on appropriate algorithm taking into account the ranking of the journal from which the meta-data was obtained as well as the number of times a certain article has been cited. This should give some estimate for the credibility of the published data. In parallel, the documents will be categorized by their appropriateness for the different user communities identified. Additionally, for a deeper evaluation of the data base, a large subset of full PDF articles will be selected from meta-data; which will form a closed data base for advanced data mining. The results of such search will be posted in the evaluation section of the data base. The data mining tools for carrying out such an evaluation will be developed in WP1 and will be validated by WP2. Validation at each stage of the development of the data mining tools (quality control) will be performed by comparing the automated data mining tools with human information extraction. This task will be executed through iterative procedure to optimize the quality of data extraction. Finally, all the activities related to the creation and maintenance of tools and infrastructure - including hardware, software and networking services - are the domain of WP3 whose goal is to establish an operational environment that will meet the long-term required quality of service.
The following figure schematically describes the planned system architecture.
In parallel to the other WPs' tasks, WP 4 activities will concentrate on identifying relevant stakeholder groups (also international bodies such as: OECD working Party on Manufactured Nano-Materials, ISO and CEN, the different EC DG's, etc.) and on gathering information related to their specific interests for making more efficient the design and content of the NHECD data base. Two survey questionnaires will be administered to a large sample of the targeted population for the data base to find out about their needs and, afterwards, to learn about their satisfaction level regarding the features and functionality of the data bank.
WP4 activities will also focus on spreading information related to the project to the different stakeholder groups. Main instruments will be: a public website, a networking meeting with the coordinators of relevant FP6 and FP7 projects, data base providers and publishers in order to present NHECD and to establish collaborations, a printed leaflet about the project which will be distributed to interested parties, and presentations at relevant events.
The overall work plan will be overseen and managed by the partners involved in WP0 (TAU and tp21).
The expected result is the creation of long-lasting data base which will serve:
