A New Evaluation Model to Building Materials Science Domain-Specific Search Engine

Materials science domain-specific search engine are becoming increasingly important because they offer increased accuracy and extra characteristics not applied by general, web-wide search engine. The current tools provide little support for materials end users. This paper proposes a new evaluation model to greatly improve the efficiency and maintenance of materials science domain-specific search engine. We describe a new model, which contains a material lexical analyzer, which pays special attentions to molecular formula and such complex data by making it easy to use and effective, and a evaluator, which evaluates the searching results. This materials data segmentation is based on molecular formula and synonym filter. A chemical element dictionary is constructed for segmenting Chinese materials name and molecular formulas. And then synonym filter is built by adding synonyms for searching relevant documents. The evaluator is built by applying different weights according to the materials science data. Our evaluations show that the model in search engine is able to produce more precise answers to user queries in materials science domain.