Total Recall is an important task in supervised text mining. The objective of this task is to find everything about an entity or about a research question. For example, in Technology assisted review (TAR), the objective is to find all the documents relevant to a request for production in a legal matter, whereas, Systematic Review (SR) involves formulating a research question, searching in multiple biomedical databases, identifying relevant Randomized Control Trails (RCTs) based on abstracts and titles (abstract screening) and then based on full texts of a subset thereof, assessing their methodological qualities, extracting different data elements and synthesize them, and finally reporting the conclusions on a particular review question. The first step in this process is to browse a large collection of abstracts and label them as relevant or non-relevant for the given research question. However, abstract screening is a cumbersome process, and thus a sophisticated methods to solve the problem is of great importance.

Publications

[1] Tanay Kumar Saha, Mohammad Al Hasan, Chandler Burgess, M. A. Habib, and J. Johnson. “Batch-mode active learning for technology-assisted review.” 2015 IEEE International Conference on Big Data (Big Data), (pp. 1134-1143). [PDF]

[2] Tanay Kumar Saha, Mourad Ouzzani, Hossam Hammady, A. K. Elmagarmid. “A large scale study of SVM based methods for abstract screening in systematic reviews.” [PDF]

Patent Application

[1] Johnson, Jeffrey A., Md Ahsan Habib, Chandler L. Burgess, Tanay Kumar Saha, and Mohammad Al Hasan. “Apparatus and Method of Implementing Batch-Mode Active Learning for Technology-Assisted Review of Documents.” U.S. Patent Application ¹⁵⁄₂₆₀₄₄₄, filed September 9, 2016.

[2] Johnson, Jeffrey A., Md Ahsan Habib, Chandler L. Burgess, Tanay Kumar Saha, and Mohammad Al Hasan. “Apparatus and Method of Implementing Enhanced Batch-Mode Active Learning for Technology-Assisted Review of Documents.” U.S. Patent Application ¹⁵⁄₂₆₀₅₃₈, filed September 9, 2016.