SLANGZY: a fuzzy logic-based algorithm for English slangmeaning selection

摘要

The text present on online forums and social media platforms conventionally does not follow a standard sentence structure and uses words that are commonly termed as slang or Internet language. Online text mining involves a surfeit of slang words; however, there is a distinct lack of reliable resources available to find accurate meanings of these words.We aim to bridge this gap by introducing SLANGZY, a fuzzy logic-based algorithm for English slang meaning selection which uses a mathematical factor termed as "slang factor" to judge the accuracy of slang word definitions found in Urban Dictionary, the largest Slang Dictionary on the Internet. This slang factor is used to rank definitions of English slang words retrieved from over 4 million unique words on popular social media platforms such as Twitter, YouTube and Reddit. We investigate the usefulness of SLANGZY over Urban Dictionary to find meanings of slang words in social media text and achieve encouraging results due to recognizing the importance of multiple criteria in the calculation of slang factor in the algorithm over successive experiments. The performance of SLANGZY with optimum weights for each criterion is further assessed using the accuracy, error rate, F-Score as well as a difference factor for English slang word definitions. To further illustrate the results, a web portal is created to display the contents of the Slang Dictionary consisting of definitions ranked according to the calculated slang factors.