The TextAnalyst package was easier to start working with than any of the others. Its user interface was relatively simple and it came with a tutorial which showed how to get it up and running in a short time. It did a reasonably good job of deriving the key points of a given text. Its abstracting capabilities managed to get the gist of the article but id didn’t grammatically clean up the results the way the IBM pakage does. It also did not have the ability to save its results automatically – you have to use the Windows cut and paste features in order to save the feature lists or abstracts that TextAnalyst generates. It is definitely intended for interactive use rather than batch processing. Its speed was quite good; I never had to wait more than a few seconds even when processing documents twenty pages or longer. It also showed a good ability to work with many documents at a time so you can derive the meaning of a whole collection of documents rather than just one. The role of this software would be to help someone digest, index and summarize a number of documents, which had already been gathered, in interactive mode. The new version which is supposed to come out any day now promises many enhancements over the current version.

TextAnalyst is an intelligent text mining and semantic information search system. TextAnalyst implements a unique neural network technology for structural processing of texts written in natural language. This technology automates the work with large volumes of textual information and can be applied effectively to perform the following tasks:

• Creation of knowledge bases expressed in a natural language for creation of hypertext and expert systems
• Automated creation of a hierarchic structure of topics for the explored text
• Creation of an adjustable semantic net of the investigated text
• Automated indexing of texts with the creation of hypertext representation
• Semantic search for information in texts
• Automated abstracting of texts
• Automated selection of the main concepts of arbitrary texts (words and word combinations), as well as semantic relations between these concepts, with assessing the relative importance of these relations

In TextAnalyst a semantic network is built automatically on the basis of an advanced homogeneous text processing. The user does not need to specify any rules or provide seeds for developing a network. A created semantic network depends only on the structure, vocabulary, and volume of the analyzed text. Megaputer claims that TextAnalyst implements algorithms similar to those used for text analysis in the human brain.

You need TextAnalyst if you often have to work through large volumes of textual information (books, articles, reports, summaries, technical documentation, etc.) want to create and update automatically a knowledge base in your subject field based on the analysis of available set of texts are interested not in a formal but in semantic search for information in texts just wish to have on your PC a powerful and convenient tool, which allows you in the course of minutes become familiar with the main issues of arbitrary English or Russian texts of a megabyte volume, or turn these texts into hypertext by pushing a single button.

James Lawson
The McIntire School of Commerce
Virginia University