The authors discuss the development of a Software Toolkit for Analysis Research (STAR) that will facilitate the development and interchange of algorithms for locating phenomena of interest in large quantities of data. Using this toolkit, researchers will be able to ascertain which existing techniques are the most promising, develop new and possibly more effective methods, and add/delete algorithms without major re-design work. This report describes the components of STAR, how these components will work together, and the algorithms with which they currently have expertise. Some modules or components of STAR will preprocess incoming data; some will select which information is appropriate for a particular application; some will analyze data to uncover items of significance; and others will assess the effectiveness of the various components. Some of the specific techniques employed by the various modules will be feature selection algorithms, machine learning algorithms, a purely statistical model, and expert system methodologies.