| Summary: | 碩士 === 國立暨南國際大學 === 資訊工程學系 === 95 === In language learning, vocabulary acquisition is basic and dominates the efficiency of the language learning. The most efficient way to increase the vocabulary is through reading which can help learner to assimilate large amounts of words. However, how does a learner understand an unfamiliar word while in reading? Traditional way is to consult a dictionary to get the meaning and then recite it frequently. The monotonous way is not very efficient. Hence, many researchers develop different methods to help users on vocabulary acquisition.
The purpose of this thesis is to develop a system different from traditional way to help users increase their vocabulary. We want to utilize the stimulation of visual sensory to impress the users about words, especially in tangible nouns. The system uses the resources on Internet to obtain the needed information, for example, getting related images about a word through Google Image Search. However, due to the diversity of documents on Internet, the search results with topmost ranks may not match the interpretation of the queried word in the page. Therefore, our system is designed to solve this problem.
The meaning of a word in context should be related to the content. Thus, our system wants to extract the representative words in the context. These words are treated as collocation of the queried word and they are combined with the queried word to form a new request to the search engine. We think this way can control the search results under related scopes. In our thesis, we describe two strategies to obtain collocations.
Through our system, users will have experiences different from traditional way on consulting vocabulary. This is helpful to impress users and to vocabulary learning.
|
|---|
