Comparative Sentences Analysis

2096 Words9 Pages

Comparative sentences are those in which the writers compare the product being reviewed with some other product. Comparative sentences are resolved by this system. Reviews with comparative sentences can be dealt with using the Named Entity Recognition (NER) method. Using this method we can assign the sentiment of the user towards this product to be the same as the sentiment towards that product. This will thus improve the accuracy of the aspect mining system as more opinions and their corresponding aspects will be considered. RiTa.Wordnet can be made use of to assign polarities to opinions and then opinion aggregation is done to identify the overall sentiment of the users towards each aspect of the product. The advantages of this system is that it resolves comparative sentences, …show more content…

Stanford POS Tagger is the tool used for part-of-speech tagging .The output of the POS Tagger will be a tree of the form, (ROOT (S (NP (DT The) (NN Cadillac)) (VP (VBZ is) (NP (NP (DT a) (JJ wonderful) (NN car.It)) (SBAR (S (VP (VBZ gives) (NP (JJ great) (NN mileage))))))) (. .))) With this tree we can identify both the part-of-speech of a word as well as22 the grammatical relation that is present in the sentence. With this information identification of aspects and their corresponding opinion is done. The above output of the POS Tagger in the form of a tree is given below pictorially. Fig 5.7 Phrase structure tree 5.2 ASPECT AND OPINION EXTRACTION 5.2.1 ASPECT IDENTIFICATION AND OPINION EXTRACTION After the parts-of-speech in a sentence along with the grammatical relation have been identified aspects present in the sentence and their23 corresponding opinions are recognized. Every product that we buy has different aspects or features which can be commented upon individually. For example, consider cars to be the product, the features of a car include its brakes, suspension, spaciousness, seats, etc. and all of these can be reviewed by the

More about Comparative Sentences Analysis

Open Document