Seminar Topics - Electrical - Electronics - Mechanical - Civil - Chemical - Computer - Automobile | |||||||||||||||||||||||||
SEMINAR TOPICSNEW ENGINEERING SEMINAR TOPICS WITH ABSTRACTS | |||||||||||||||||||||||||
|
Context Disambiguation on Web Search Results
Among the most important intentions of web usage is information retrieval and the most common activity pursued to achieve that is web searching. As the web contains information on virtually all topics, precision is a very important yardstick to measure the quality of search engines. This report describes a method to make web search results more customized in certain cases where the query terms have multiple senses or even referents.
Here first the problem at hand and the nature of the proposed solution is defined. Then we deal with the different issues that concern such a solution along with the reviews on some related works of interest. Then our approach is described plus the results of experiments using this approach. The last section lists the conclusions.
It can be seen from the tests that the algorithms perform very well even when they have very little information to work with. Thus implementing context/referent disambiguation as a meta-service may be an interesting experiment. The obvious concern would be as to how the algorithm would scale with increase in the number of pages in input. Search engines usually provide results in pages of 10 (or more) results. Thus, a meta-service implementation of the disambiguator need process only 2-3 result lists, and thus 20-30 pages, at a given time and can defer the disambiguation of other pages to a time when they are actually required. Thus we are justified in ignoring the problem of scalability altogether if the algorithm is to be implemented as a meta-service. But in cases where the algorithm is to be embedded into the search algorithm, such issues may assume greater significance.
Future work may be directed towards formulating possible optimizations when the disambiguator is implemented within a web search algorithm (in the search engine architecture itself). The optimizations may be formulated to exploit the larger corpus. Further link based information may also be used in such cases. The keyword identification algorithm can also be improved considerably making use of principles and results from computational linguistics. The performance of the algorithms in very broad topic queries such as “union” whose context ranges from “rugby unions” to “c unions” and in cases of word sense ambiguities and other issues, not dealt with in the current paper, have to be investigated. Such cases are clearly different from the cases discussed through the bulk of the paper, but investigations as to how the algorithms presented here work on such problems, might provide valuable clues on what the nature of solutions for such varied problems. | ||||||||||||||||||||||||
| FIND YOUR JOB IN INDIA |
| ||||||||||||||||||||||||
|
|
| |||||||||||||||||||||||
