Text mining of medical/biological abstract

With the dissemination of microarray technology, scientists hope to get a broader understanding of functional meanings of and the relationships between different genes. To date the mainstream data analysis strategies to do so, consist of clustering expression data in combination with the tedious task of manual functional inference.
In recent pioneer's work, the use of electronic literature as a supportive methodology for functional inferences has been proven to be a viable path. We are currently investigating an extension of the power of text mining in the area of post-genome informatics by integrating electronic document management and statistical data analysis. In a first phase, we wish to reveal the correspondence between the textual world (e-articles) and the data world (gene expression data) by assessing the validity of clusters generated by statistical data mining algorithms in both domains.
In a second phase, we are looking how to integrate the clustering process in both domains and how this can enhance the process of accurate knowledge discovery.

Back to top