Aspect-Based Opinion Mining

Create a training data set for opinion mining The main aim is know what is the trend that the data is following. For example: We need to know where is the location of the feature or the aspect of the product is located in the sentence and the same for opinion as well, we need to know where is the location of the opinion in the sentence using some linguistic and data mining to find the rules that the data is following . Previously the location of the opinion is the closes adjective to the feature/ aspect , ((close to the feature (check [login to view URL]~liub/FBS/[login to view URL] and it is not right all the time)), however we aim to know what are the rules that the data is following to determine the location of the opinion and the feature , how far is the feature from the opinion. sometimes the orientation is not right and finally the feature and the opinion may be extracted, however the extracted opinion may not represent the right feature and vies versa. For example: Orientation Feature Opinion Sentence 1 Camera (noun) Easy ( opinion) the camera is very easy to use ; in fact on a recent trip this past week i was asked to take a picture of a vacationing elderly group . 1 Use ( the feature is consider to be verb) easy the camera is very easy to use ; in fact on a recent trip this past week i was asked to take a picture of a vacationing elderly group . -1 camera Smaller ( this opinion was expressed on other feature) comparison to other product !!! therefore the opinion was wrong and the orientation is wrong as well) it supposed to present a positive opinion , where is extracted -1!! the g3 is loaded with many useful features ; and unlike many smaller digital cameras ; it is easy to hold steady when using slower shutter speeds . -1 camera Easy ( same as previous). the g3 is loaded with many useful features ; and unlike many smaller digital cameras ; it is easy to hold steady when using slower shutter speeds . So we have set of sentences { S1 , S2 .... S n}, we need to extract features and opinions, orientation along with the location of (the feature and the right opinion) in the sentence. 1. Features and their location in the sentence using [login to view URL] [login to view URL]~lzhang3/paper/[login to view URL] POS tagging and bag of words 2. Opinions and their location in the sentences as well using POS tagging , chucking and dictionaries [login to view URL] [login to view URL] [login to view URL]:_A_Publicly_Available_Lexical_Resource_for_Opinion_Mining [login to view URL] [login to view URL]~liub/FBS/sentiment-analysis.html#lexicon consider that the opinion are not always adjective it may be any other part of speech. 3. Find the orientation ( positive , negative , neutral ) 4. Then map the relation between features and opinion in pairs { (f1, 3) ,(o8, 8)} for each sentence 5. Then have list of rules that the data follows , you may use association rules. Background information: 6. Bing liu web page [login to view URL]~liub/FBS/[login to view URL] 7. [login to view URL]~lzhang3/paper/[login to view URL] 8. POS tagging [login to view URL] 9. Sentence detection [login to view URL] 10. Text and Data mining ( Google) 11. Opinion mining ( Google) Data Set: [login to view URL]~liub/FBS/[login to view URL]

C# Programming Software Architecture

Project ID: #5123854

About the project