| Summary: | The current standard way of searching for information is through the usage of some kind of search engine. Even though there has been progress, it still is mainly based on the retrieval of a list of documents in which the words you searched for appear. Since the users goal is to find an answer to a question, having to look through multiple documents hoping that one of them have the information they are looking for is not very efficient. The aim of this thesis is to improve that process of searching for information, in this case of medical knowledge in two different ways, the first one is replacing the usual keywords used in search engines for something that is more natural to humans, a question in its natural form. The second one is to make use of the additional information that is present in a question format to provide the user an answer for that same question instead of a list of documents where those keywords are present. Since social media are the place where people replace the queries used on a search engine for questions that are usually answered by humans, it seems the natural place to look for the questions that we aim to provide with automatic answers. The first step to provide an answer to those questions will be to classify them in order to find what kind of information should be present in its answer. The second step is to identify the keywords that would be present if this was to be searched through the currently standard way. Having the keywords identified and knowing what kind of information the question aims to retrieve, it is now possible to map it into a query format and retrieve the information needed to provide an answer.
|
|---|