Explore Courses Blog Tutorials Interview Questions
0 votes
in AI and Deep Learning by (50.2k points)

My CSE graduation project I chose to be a simulation of a search engine that uses sentiment analysis to evaluate whether comments/reviews is positive/negative/neutral

I am not sure how would I be doing this yet, but I understood that it uses classifying agents that you train first then it can classify by itself. the edge is not high (about 60% I understand)

My question is, will that work on non-English languages? I mean I want to do sentiment analysis on Arabic comments/reviews and generate a proper sentiment, is it doable with the current classifiers? and if Not, I would need to write my own classifier, is it possible to write a custom classifier?

1 Answer

0 votes
by (108k points)

One of the central problems in multilingual sentiment analysis is a significant lack of resources. Thus, sentiment analysis in multiple languages is often presented by transferring knowledge from resource-rich to resource-poor languages, because there are no resources available in other languages. The majority of multilingual sentiment analysis systems use English lexical resources such as SentiWordNet.

Another approach is to use a machine translation system to translate texts in other languages into English: the text is translated from the original language into English, and then English-language resources such as SentiWordNet are employed. Translation systems have various problems, such as sparseness and noise in the data. Often the translation system is not able to translate necessary parts of a text, which can cause serious problems, possibly reducing well-formed sentences to fragments.

For more information regarding the Multilingual Sentiment Analysis, refer to the following link:

If you want to know how to do sentiment analysis then visit this Machine Learning Course.

Browse Categories