Probabilistic search service based on user history on cloud

People are continuously pursuing intricate task-oriented goals on the web, such as purchase planning or managing finances or even buying digital files. To this, the user conventionally breaks down the tasks into a few co-dependent steps and issues multiple queries around these steps perpetually over long periods. To help users in their long-term information quests on the web, search engines maintain their queries and clicks while probing online. In order to give the search results quickly a shallow search will be made on the user history so that the results will be populated quickly. Pruning the search results can be done by various factors. The user will look for data that is more relevant to where the user is living. Another way would be to consider the most accessed contents. It analyses another approach to tackle the “tiny target space” problem. This is based on the assumption that the user would never know that an item exists in the system if it is no way connected to the user. If the user gets to know the existence of some item, all those relevant to this item might interest the user. The recommendation space thus being smaller than the actual search space, it would be possible to search contents that are relevant to the user, but not all. So this can be used to show preliminary results to the user by the time the system fetches the entire results for the user. This would reduce the idle time the user spends on the system. This paper analyses various techniques to be used to deploy such a search service that reduces the time taken to produce the first search result.