Honmonymy is when a word has multiple(entirely different) meanings. For example,cosider the word ‘pupil’. It can either refer to students or eye pupils depending on the context in which it is used.
Suppose you are searching for the word ‘pupil’.The seach engines sometimes give data relevant to the context and some times give irrelevant data.Such things happen because the query may contain words whose meaning is ambiguous or those having several possible meanings.
To solve this ambiguity problem, the Lesk algorithm is used.
In the previous segment, we understood different functions in WordNet. In the next video, you will learn how WordNet can be used to disambiguate words.
You can consider the definition corresponding to the different senses of the ambiguous word and determine the definition that overlaps the maximum with the neighbouring words of the ambiguous word. The sense that has the maximum overlap with the surrounding words is then chosen as the ‘correct sense’.
“She booked the flight tickets to Delhi in advance”
In this example, you will notice that the gloss or meaning1 has more overlap than gloss2 Hence,we consider gloss 1as the correct sense.
Although we have shown only two senses,a lesk algorithm parses through all the word senses and outputs the gloss that has the maximum overlap.
In the next segment, you will learn how to understand how to use WordNet to code the Lesk algorithm in NLTK.