| With the expeditious increase of biomedicine data, it not only accelerates human's intercommunion and research, but also makes people to face the immensity data at loose ends, So there comes three raring requirements: how to expediently search information of gene regular field? How to know the newest investigation intime? And how to find papers fleetly? At the same time, data mining technologies become popular, especially Text mining. these technologies also make it possible for us to solve the aforementioned three questions.this topic puts forward a new words mining method based on frequent sequence arithmetic, develops frequent sequence arithmetic system (FSAS) to serve for paper selection and text classification. Simultaneity, with the help of text mining technology, we create the Gene Regulation Information Source DataBase (GRIS) and Gene regulaion oriented information agent system (InfoAgent).Text classification is the main part of text mining, and it's kernel is how to get text characteristic. The traditional word segmentation technique is no more fit the requirement of gene regulation research work. FSAS could extract long-words and new-words from texts without wordbook。These words extracted with FSAS generally represent the main idea of a paper ,so they can help people to know the main idea of papers, and to choose right paper which they really need. These words can also be used to build text characteristic, and to work for classify texts. We use the text eigenvectors extracted with FSAS and SVM technique to do some experiments, the average accuracy is above 85%. In addition, with a number of gene regulation texts dealed with FSAS, people can get the professional keywords of this field.To create GRIS, we collect the correlative information of gene regulation in internet, such as DataBase,tools and literature . People can handily get all kinds of gene regular information resource.How to choose the literatures from the vast gene regulation literatures has...
|
PDF Full Text Request