@techreport{TD:100896,
	att_abstract={{Social media contains many types of information which is useful to businesses. In this paper we discuss automatic extraction from twitter data the descriptions of problems consumer experience with products and services. We first identify the problem tweets i.e. the tweets containing descriptions of problems. We the extract the phrases that describe the problem. In our approach such descriptions are extracted as a combination of trigger and target phrases. Triggers are mostly verb phrases and are identified by using hand crafted lexical and syntactic patterns. Targets on the other hand are noun phrases related to the triggers. We frame the problem of finding target phrases corresponding to a trigger phrase as a ranking problem and show the results of our experiments with maximum entropy classifier and voted perceptron. Both approaches outperform the rule based approach reported before. We also show that because of inherent limitations of voted perceptron, maximum entropy based ranking are more suitable for our problem.}},
	att_authors={ng2836},
	att_categories={C_CCF.2, A_ST.2, C_IIS.2, C_IIS.11},
	att_copyright={{}},
	att_copyright_notice={{}},
	att_donotupload={},
	att_private={false},
	att_projects={OpinionMining},
	att_tags={Information Extraction, Socail Media, Twitter, Natural Language Processing},
	att_techdoc={true},
	att_techdoc_key={TD:100896},
	att_url={http://web1.research.att.com:81/techdocs_downloads/TD:100896_DS1_2012-05-24T17:05:41.856Z.pdf},
	author={Narendra Gupta},
	institution={{Conference on Intelligent Text Processing and Computational Linguistics
CICling 2013}},
	month={March},
	title={{Extracting phrases describing problems with products and services from twitter messages}},
	year=2013,
}