Power Text Solutions

Text Mining by Power Text Solutions

We offer two independent approaches to information summarization and analysis:
A. iResearch Reporter is recommended when you need to quickly familiarize yourself with different topics.
B. Topical Research Information Modules (TRIM) are good when you have permanent research areas.

TRIMs can be viewed as specialized semantic modules, tuned up to your topic. They provide multiple choices for information processing and in-depth analysis. TRIMs structure the data according to the automatically identified key aspects of user's research theme, and hence provide a multi-perspective, panoramic view of information in your research area. Each TRIM can be reused repeatedly - just load it with a fresh portion of documents.
Read more on the TRIM approach and evaluate this system here.

iResearch Reporter multi-document summarization technology allows to summarize web search results on the fly, thus saving your time by performing automatic research job at your request. Instead of directly accessing the relevant documents provided by the search engines, users get the well-structured pertinent information extracted from these documents.

Besides using results from the common web search, this technology can also be applied to processing of information retrieved from search engines (e.g., Google, Yahoo, Google Scholar, etc), online databases, web forums & blogs, specialized search engines, etc. It also can use custom lists of URLs or summarize collections of documents right from your desktop.
Along with the multi-document summarization, our tools also support non-query-biased summarization of individual documents. Summarization can take into account multiple general (like Trends & Forecasts, Advantages & Disadvantages, Causes & Consequences, etc, etc) or user-defined (professional) contexts, thus focusing the output onto particular perspectives and organizing it into relevant thematic sections.
Read more on the iResearch Reporter and evaluate this system here.

At the core of our technology is methodology of excerpting the most relevant and informative text passages that are semantically and grammatically complete. The measure of such passages provides effective criteria for smart ranking of documents by their relative pertinence. Such text passages are used for compiling self-sufficient overviews of user-described topic. These comprehensive automatic digests can be used as a preview or even instead the original documents. The technology can also be applied for summarization without a user query of an aggregated set of documents relevant to a certain topic (e.g., articles provided by the Google News).

Our automatically compiled multi-document summaries reach the quality of human-written overviews (independently of the amount of source documents and with any length of the output text). The summaries are characterized by meaningful organization, i.e. they have contents, sections, subsections with subheadings, paragraphs. Our approach allows to modify the summarization techniques after guessing the user's search intentions through analysis of a query formulated in natural language form.
Our NewsFeed Researcher public project provides continuous summarization of Google News topics in the business, technology, US, international, entertainment and sports areas. Compiled articles are published online with no human touch. The capability to summarize the news on any particular topic can be tested "live" by evaluating the iResearch Reporter advanced demo mentioned above.

Important feature of this technology is its high flexibility, allowing to quickly adjust it to be able to retrieve and organize information under a certain angle, e.g. applying certain general context or tailor for a specific professional area (like emergency response and law enforcement, homeland security, real-estate, medicine, journalism, different aspects of business research, etc). This allows us to quickly deliver custom solutions for different professional areas, accommodate to special user tasks and different kinds of information resources.

We offer free and commercial web services, so as building custom applications for information professionals.

Information page devoted to creator of this original technology, Martin Soubbotin, is currently available in Russian.

Java courses