Septet offers the following innovative products:
Plagium
Plagium (www.plagium.com) is a Web-based tool that assists in the discovery of plagiarized content on the World Wide Web. Likewise, Plagium can be applied to track the legitimate or fraudulent usage of content, such as syndicated news stories carried by multiple Web sites. Plagium can also be used to track the appearance of fraudulent Web sites whose purpose is to steal user confidential information (such as credit card information), while attempting to resemble a legitimate Web site.
Plagium is offered free of charge to casual users. A Deep Search capability enables users to perform a more exhaustive analysis of larger documents for usage of text within the documents over the Web. Plagium's alert feature automatically warns users on a periodic basis when a newly found Web site contains text of interest.
Septet is now working on API modules that will enable Web site owners and text management services to incorporate an in-line plagiarism checking facility.
TX Miner
TX Miner is a sophisticated information retrieval tool that provides the underlying support for the Septet Personal Search Syndication™. In addition, TX Miner possess the following capabilities:
1) Agent-based approach to information search and retrieval. Here “agents” operate as independent processes that are programmed by the user to search and retrieve information from distributed repositories of publicly available or private information. The agents can be programmed to run at scheduled times, continually, or on demand. On-demand operation may be manual or through a triggering mechanism (e.g., in response to a news story on a specified subject matter). In addition, the agents may operate as processes on remote computing platforms, especially those platforms in proximity to their respective information sources.
Powerful analytic capabilities for categorization of incoming data, as defined by a combination of agent filter settings, subdirectory filter settings, and TX Miner’s extensive linguist analysis.
Information sharing among peers. This enables, for example, the construction of databases of highly targeted subject matter. Multiple, but qualified users, comprising a community of interest could access and refine the TX Miner database to improve search results on a topic.
2) Entity extraction, to ascertain the relationships between people, places, and time as viewed within a single document or over a large collection of documents.
3) Automatic generation of timelines corresponding to events, people, pl aces, etc. This is very useful for tracking industry trends, corporate image, actions of entities, news stories, etc.
4) Connections to popular Web search tools, such as Google, Bing, Yahoo, and others. These connections support TX Miner’s meta-search capabilities. But unlike conventional meta-search engines, search results are gathered through agents and processed through TX Miner’s extensive analytic capabilities.
Generation of email alerts to specified individuals or communities in response to user-programmed agent search / filter results.
5) Publication of category-based document lists through automatically generated RSS lists. These lists can then be retrieved as links placed into popular RSS readers for easy extraction and listing of updated, topic-specific web (including blog) documents.
6) Search capabilities over topic-specific groupings of documents from various sources. This enables the creation of specialized search engines tuned to specific subject matter. These search engines can also be constructed through TX Miner’s peer-to-peer search refinement capabilities.
TX Miner operates in conjunction with MySQL or Microsoft SQL Server databases operating on Windows 2003, Windows Server 2007, or Vista. Current Web delivery services are via Cold Fusion MX. A Linux version of TX Miner is presently under development with all Web services to be controlled via Java. Operation over clustered servers is available upon request. Web access through latest versions of Internet Explorer, Netscape, and Firefox is presently operational.
K-Sync
K-Sync is a web-based application that enables groups of users, even the general public, to create and share organized sets of information. Through its social networking-like functions, groups of users can collectively define the information sets, which are populated by automatically retrieving relevant documents and pages from the World Wide Web or corporate databases. This document retrieval makes use of an proprietary agent-based filtering system that assesses the content for names, times, places, and other characteristics.
The open architecture of our award wining k-sync platform enables the incorporation of many of k-sync features into your enterprise intranet or even public internet. Thanks to the use of standard XML, Java, and other easy-to-integrate technologies, Septet now offers these functions through our k-sync API library:
1) Entity Extraction. Automatically "extract" proper names, places and dates from documents.
2) Categorization and Classification. The powerful and proven categorization capabilities of TX-Miner are now available to any Web system. APIs make use of user-specified categorizes through boolean descriptions. Our algorithms will classify documents from there.
3) Relationship Extraction. Enables the listing and graphical display of cross relationships between entities within one or a corpus of documents.
4) Concept extraction. IPTC ( International Press Telecommunications Council) onthology compatible, with implementation based on English and French Dictionaries.
5) Septet Meta Search Engine enables search over multiple search engines.
Contact Septet to learn about how our APIs can work for your enterprise information management or even public Web application.