From rxpgnews.com
New search engine evaluates user's personality
By Jatindra Dash, RxPG News Service,
Dec 10, 2006 - 4:23:48 PM
New Delhi, Dec 10 - A group of Indian researchers claim to have developed an Internet search technology that can recognise the personality of a user and automatically align the results according to his or her interest.
The technology uses proprietary algorithms - a privately owned method to find a 'personality' of the user based on tensor algebra - a branch of mathematics mostly used by a physicist in studying Einstein's theory of relativity, Sanjeev Sanghi of the Indian Institute of Technology - Delhi told IANS. 'The analysis is much more than a mere keyword search or counting,' he said.
It can be used at any place where an automated ranking is needed based on the personalized concept search, Sanghi, who heads Tensor Technologies Pvt. Ltd, a Delhi-based company, said.
One can evaluate the technology, which is posted for validation and upgradation by Beta users at http://search.t6labs.com, he said. It asks a new user to first register. After this the user can start searching by giving any query on the web, he said.
'Once he gets the results, he will click few pages of his interest. The site, then, starts to understand the interests of the user based on the content analysis of the pages which have been clicked by the user'.
'After this when the user fires a new query, the site will automatically align the search results according to the interests of the user. A few key features of the site include the more you use the site, the more it understands you.
It works in real time, quickly adapts to user's interests and delivers results as per user's interests. It does not ask for any explicit or fully and clearly expressed or demonstrated user input.
'For example there is a user 'john' with password 'john' who has browsed 3-4 sports sites. The browsed sites which count in the personality evaluation can be seen from the site itself,' Sanghi explained.
'Now when this user gives a title 'tigers' the user will find that the top 4-5 results are for tigers baseball team and also for a tigers rugby team. The sites with tiger as an animal or tiger as a software are ranked lower than the tiger sports sites.
'In contrast, another user who has browsed pages on animals like lion or cat will find the web pages related to animal tiger ranked higher than the pages containing tiger software or Tiger Woods.'
'The idea arose while Chandra Shekhar -, an alumnus of IIT Delhi majoring in maths and computing, was doing his Masters project with me,' Sanghi said.
'We joined hands with Shashank Khare, a computer science graduate from IIT Kharagpur and developed this.
'There is no robust technology available over web for generic purpose use. One with Google asks you to wait for a long time to see any kind of personalised search results.
'The technology with Yahoo for personalisation is a bi-polar search i.e., for each query it tells you whether you'll like to get results according to scientific research interest or by business interest.
'The other sites which claim personalisation take specific and explicit input from the user. Our technology 'deciphers' your interests automatically on the basis of the pages you visit,' Sanghi said.
'As we get user interest captured and we deliver user specific contents one commercial application will be user specific advertisements that will lead to higher click conversion for ads,' he said.
'It can also be very useful for personalised product recommendation i.e., user specific books, movies etc. The commercial implications are enormous.
'We have already filed a patent with the US Patent and Trademark Office and are in the process of filing other patents too.'
The technology can be used in automated product recommendation services and includes a wide variety of applications i.e., recommendation of books, movies, holiday package, hotels etc.
The technology will be applicable to any situation in which there is a bunch of documents that have to be ranked. This will rank the documents according to each individual user's interests, Chandra Shekhar said.
Another place where this technology can help is in finding users who have similar interests. This is of use in social networking where the technology will help the user to recognise other likeminded users.
Commercially this could then be exploited to find experts who could answer the queries by other users, may be on a payment basis.
'With our technology when a user gives a new search then his/her personality, which has been dynamically evaluated, gets matched automatically with the available documents on the topic and the results are shown in new order,' he added.
All rights reserved by RxPG Medical Solutions Private Limited ( www.rxpgnews.com )
|
|