Literature and Text Mining

Is data quality acceptable?

Is data generated from validated methods?

Are results validated?

Biological data curation or bio-curation is a process of collecting and organizing biological data. The goal of curation is to create structured, comprehensive, integrated and accurate source of biological knowledge.

Over the years, the data type and quantity has been changing as the number of publications have increased exponentially. Curation of this huge number of publications is daunting and it requires trained and well qualified personnel in the field and new curation tools that improve speed and accuracy.

Apart from data curation, we provide other value added options such as scoring or ranking system for any specific field(s) of interest and an option to validate the curated data. Scoring or ranking system determines completeness or importance of the data that is retrieved, and also it enables sorting of the results. Providing an option to user to validate the curated data increases the confidence and correctness on the curated data in the subsequent searches by the other users or the same user.

We use semantic web search engine for data mining for complete coverage of literature. For data extraction, we use in-house information extraction system.

Challenges:

1. Heterogeneous data: Data in present in different databases, non standard formats and data is generated from different platforms.
2. Valid Information: Is data quality acceptable? Is data generated from validated methods? Are results validated?
3. Data Storage:Optimizing models to predict and filter the right information.

Benefits

1. Data mining and curation is handled by team of experts with specialized skills .
2. A three step data validation mechanism is used.
3. Mined data is stored in relational databases / XML for easy retrieval and updation of information
4. Customized reports can be generated to view desired data.