nlp - How could I identify a sentence disclosing some specific information in a paragraph? -


for example, have such paragraph below: first sentence (bold , italic) hope identify out. identification goal includes: 1. whether paragraph contain such disclosure. 2. disclosure is.

the possible problems : 1. sentence may not in begin of text string. in place of given paragraph. 2. sentence may vary words same meaning. example, expressed as: "sample provided review" or "they sent me item evaluation" or this.

so how identify such disclosures ? anyone's idea appreciated. thanks.

the paragraph:

i sent earbuds audiophile headphones review. going copy here information site: "high definition stereo earphones microphone equipped 2 9mm high fidelity drivers, unique sound performance, well-balanced bass, mids , trebble. designed specially enjoy classic music, rock music, pop music, or gaming superb quality sound. let cor3 in ear sports earbuds. replaceable caps, inline controller , mic extreme flexible tangle free flat tpe cable including inline controller universal microphone. play/pause music or answer/hang call touch of button right next hands, feature available depending on device capability. cor3 should best gaming earbuds. extremely comfortable

methods have tried: now, processing naive: 1) humanly labeled 1000 pieces of reviews binary variable (1 represents including disclosure text, 0 otherwise). 2) collect disclosure texts corpus denoted disclosurecor; 3) based on these disclosurecor, discovered basic regular regression rules, " review.* evaluation|test|opinion". 4) using these summarized rules label new data. 5) problem rules may not complete, since own subject summarizations. besides, theses rules may not occur in disclosure text, other parts in review paragraphs, generating lots of noises (i.e. low precision); 6) tried use classification based association rules train rules labeled data. keywords number huge, long long time needed train rule, crashed often. 7) tried compare similarity review paragraph disclosurecorp, it's difficult find threshold cut whether review paragraph contains disclosure. these efforts have tried, please give me hints ? thanks.


Comments

Popular posts from this blog

javascript - jQuery: Add class depending on URL in the best way -

caching - How to check if a url path exists in the service worker cache -

Redirect to a HTTPS version using .htaccess -