The Application Of Text Mining To Engineering Drawing

With the rapid development of both information science and Internet technologies, Information Retrieval becomes attractive in Computer Science. Text Mining, as a new born branch of Natural Language Processing (NLP) which has almost 50 years history, plays an important role in Information Retrieval. Engineering Drawings, as a significant data format in industry, are widely spread on the Internet. The traditional course of Information Retrieval is based on text data format, and nowadays the multimedia data formats also take part in the course, such as audio format, video format. However, the drawings format is not in the list. In this paper, we try to make use of Text Mining in the Information Retrieval course based on drawings format. Specifically, we will pay attention to two parts: Drawings Search and Information Hiding of Drawings.Firstly, we will take advantage of Text Information Extraction into a Search system of drawings. By constructing a prototype of the Search system of drawings, we provide several methods such as the simplest one and the complex but effective one; all of them can probably fulfill the entire system. Further, we give a detailed description of experiment with simplest method, all the results of experiments with each method are evaluated finally.Secondly, we will make use of Text Classification into Information Hiding of drawings. We propose a method of extracting the confidential part from drawings automatically by text classification. We take a simple but effective method of Text Classification, Linear Classification, to the Extraction of the confidential text. By the extraction of confidential text, we make the key text away from drawings, and keep it confidential. Finally, we discuss several method of Text Classification that would probably lead the success of the system.