As a part of the joint research project reported in parallel at the conference, we have been developing a database system similar to those common in the world data centers, such as DPC at NIFS. Instead of outsourcing the software engineering work, here we focus on the development of a free-software open-source solution. In particular, our work on an on-line electron-molecule scattering bibliographic database is reported (linux operating system, MySQL relational data management system, PHP logic layer, Apache server, and HTML presentation layer) [1]. We also report on our development of the cross-section database with on-line generation of data plots using a test data set [2]. The major issue discussed in this work is the sequence: automated data collection, data mining and database update. Automated scripts browsing abstract databases of several publishers have been developed and tested. In semi-regular time intervals, new data are searched for, downloaded, extracted and uploaded to the database automatically. The present system employs a simple keyword search to estimate abstract relevance; in future, it is designed to incorporate more sophisticated methods, such as the Learning Vector Quantization algorithm (reported in parallel by other authors).

References