Web-Observer is an automated system intended for retrieving and structuring information from open Internet sources and corporate warehouses (the latter is available in corporate license only). The system finds and extracts information that has appeared, disassembles it into required attributes and saves materials in a database in unified format.
The technologies used in Web-Observer are based on converting unstructured data into a unified format and its abilities are very extensive. All incoming information is automatically categorized and presented in a structured way to a user in categories/subcategories. Upon a client's request one can create a rubric with records of company name or brands or a rubric with the latest news and information related to a specific topic, e.g. Pharmaceutics, Finance, Politics etc. The system contains a built-in search mechanism that allows a search for the information using a set of filters and taking into consideration the morphological characteristics of different languages. A user can create personal news selections related to his points of interest using a mechanism of search agents. The system immediately alerts the user via e-mail or SMS whenever the fresh piece of news has been found. Users can look through these materials in a password-protected area of the system or in his mailbox.
Unique categorization that is formed according to a client's needs allows precise reflection of the professional activity of a company. One can name the possibility of adding a large number of information sources that a client needs in his work, flexible adjustable filters that allow delivery only online, thoroughly selected information avoiding 'information trash', automatic process of information collecting, and structuring and retrieving in order to describe briefly some basic features of the system. Language independence allows retrieval of information of the client's interest by monitoring sources from all around the world.
Collecting information of different types and processing it into a unified structure, the system can become a single universal source of information for companies and organizations. All users work within a single information field and receive information in the shortest possible time after it has appeared online. Text storage in a local database guarantees permanent access to the information while the same information might not be available from the source after a time. This option is very useful and efficient if a company's employees have often-changeable priorities or tasks.
We are currently working on new product features that would hit the bottom line of the marketing departments of large companies. They will allow:
- to estimate the sentiment of articles in mass media and Internet about company's products and the company itself ('positive', 'negative', 'neutral')
- generation of information portraits on objects of interest, time dynamics and analysis of their relations.