E-membranes to detect Users’ Behavior Patterns

Darwin-FIRST, is an Artificial Intelligence procedure to automatically retrieve intelligence diluted in huge databases. FIRST stands for Full Information Retrieval System on Thesaurus, one expert system aided by agents and Darwin stands for Distributed Agents to Retrieve the Web Intelligence, a network of FIRST nodes. FIRST algorithms and agents extract from databases their inherent intelligent skeletons throughout e-membranes. These e-membranes will allow Internet users to locate Cognitive Objects directly, like it they were exposed in supermarket stands. They behave like special high resolution magnifier “glasses” that enabled them to “see” and to retrieve those Cognitive Objects precisely. E-membranes are like real bio membranes, they have their own e-endoderm, e-mesoderm, and e-ectoderm, virtual matchmaking interfaces that separate servers from their users, cognitive offer from cognitive demand. In the same way that e-membranes could be settled and tuned up to extract inherent intelligence skeletons from databases they could be settled and tuned up to extract the people’s inherent intelligence skeleton as well. The ectoderm is enabled to detect users’ behavior patterns that at users’ side are analog to keywords at servers’ side. E-endoderm takes care of messages from e-mesoderm to server’s side and vice versa. Content and its structure evolution are controlled by special messages. E-mesoderm takes care of these special messages generation related to evolution and learning processes. In theory once adequately settled and tuned up e-membranes could be left to work autonomously. However the whole process, and the “permeability” of the e-membrane layers is supervised by a Desktop, controlled in its turn by a human being, the Chief Editor. E-membranes may work in three modes: a) extracting intelligence, for instance to build a proprietary intelligent content to attract users; b) as an i-matchmaking interface between servers and their users, continuously optimizing offer versus demand, autonomously freezing/killing unused offer and/or searching new offer for unsatisfied demand; c) at users’ behavior pattern detection mode, learning as much as possible from users interactions With these e-membranes we may build super search engines, working over the Web space in first mode first, and then creating a Web Thesaurus, the inherent intelligence skeleton of the Web. We may also recover the inherent intelligence of any huge data reservoir to facilitate its overall security and daily management. In this case we make reference to corporations that along years have nurtured huge databases without taking care of saving its history indexed by their inherent database intelligence. A prototype of this e-membrane is now working in Darwin-FIRST Demo site hosted in www.intag.org. Its procurement agent worked in the first mode to extract the inherent intelligence skeleton from the Web for a single discipline: Computing. This skeleton has the form of a Thesaurus of 53,000 keywords related to a Logical tree of 1600 branches/leaves and a set of nearly 6,000 Authorities. Now the prototype is working in modes 2 and 3. Key-words:-E-membrane, Knowledge management, Knowledge representation, Web thesaurus, Data mining, People’s pattern behavior, Agents, Behavior patterns, Search engines, Man-machine 1 The apparent chaos of huge data Reservoirs The fig. 1 depicts how i-Webs, Webs with built-in emembranes, would perform their content procurement and retrieval tasks. Black region represents the Web space that hosts more than 8,000 million documents. Search Engines, like the one depicted in blue, aid world wide users to locate what they need [16]. Being this space so vast, some ranking procedure is needed. However, the Websites competence to “rank high” is so fiery to encourage all kind of dubious ethic marketing strategies and many times unethical habits oriented to deceive the Search Engines robots that continuously browse the Web space to update their ranked lists. For this reason two equally important Websites like the ones shown may look extremely different concerning their rank: one like a bright star on the Top and the other like a pale star among millions.