Stanford Classifier


Notes:

Stanford Classifier is a general purpose classifier, which takes a set of input data and assigns each of them to one of a set of categories.  It shines working with textual data, with powerful and flexible means of generating features from character strings.  Alternatives include R, Weka or scikit-learn.

Resources:

See also:

Best Stanford NLP VideosStanford NLP & Dialog SystemsStanford Parser & Dialog SystemsStanford Tregex


Twitter sentiment analysis A Go, L Huang, R Bhayani – Entropy, 2009 – www-nlp.stanford.edu … The code has the following dependencies: 1. Stanford Classifier library: http://nlp.stanford.edu/ software/classifier.shtml 2. OpenNLP MaxEnt library: http://maxent.sourceforge.net/index.html 3. Twitter4J is an external library for parsing tweets: http://yusuke.homeip.net/twitter4j/en … Cited by 73 Related articles All 6 versions

ISM @FIRE2013 shared task on Transliterated Search DK Prabhakar, S Pal – isical.ac.in … 2459–2465. [3] Karimi, S., Scholer, F., and Turpin, A. Machine transliteration survey. ACM Computing Surveys (CSUR) 43, 3 (2011), 17:1–46. [4] Klein, D. The stanford classifier. http://http: //nlp.stanford.edu/software/classifier.shtml, 2003. Online; accessed 16-06-2014. … Related articles All 3 versions

Emotion Detection In Suicide Notes Using Maximum Entropy Classification R Wicentowski, MR Sydes – Biomedical informatics insights, 2012 – ncbi.nlm.nih.gov … 4 In this work, we made use of the freely available Stanford Classifier. … Word shape: One of the features that can be automatically generated by the Stanford Classifier is word shape. This feature is used to conflate words that “look” the same. … Cited by 2 Related articles All 17 versions

Estimation of user’s activity from tweets through tri-layer clustering model D Zhu, Y Fukazawa, J Ota – Mobile Computing and Ubiquitous …, 2014 – ieeexplore.ieee.org … B. Control Group To verify the performance of the proposed model on activity estimation, the Stanford Classifier [24] was adopted as the control group. … change with the number of activities by fixing the number of topics, but the TLCG always outperforms the Stanford Classifier. … Cited by 1 Related articles All 4 versions

Analysing market sentiment in financial news using lexical approach T Li Im, P Wai San, C Kim On, R Alfred… – Open Systems (ICOS) …, 2013 – ieeexplore.ieee.org … For the automatic approach, they developed a Java data processing code and used the Stanford Classifier to analyze and predict the sentiment of the financial news articles. The Stanford Classifier utilized Maximum Entropy … Related articles

Twitter sentiment classification using distant supervision A Go, R Bhayani, L Huang – CS224N Project Report, Stanford, 2009 – s3.eddieoz.com … accessing tweets by query term. The Twitter 3The Stanford Classifier can be downloaded from http://nlp.stanford.edu/software/classifier.shtml. 4More information about the Twitter API can be found at http://apiwiki.twitter.com/. Page 4. … Cited by 594 Related articles All 5 versions

Sentiment analysis of occupy wall street tweets R Chang, S Pimentel, A Svistunov – 2011 – Citeseer … All test errors were computed using leave-one-out cross-validation except for the Stanford Classifier’s. … Retrieved December 10, 2011 from http://nlp.stanford.edu/software/classifier.shtml [7] F. Å Nielsen, AFINN-111 (list of valence-rated English words), Informatics and … Cited by 1 Related articles All 2 versions

JMaxAlign: A Maximum Entropy Parallel Sentence Alignment Tool. M Kaufmann – COLING (Demos), 2012 – aclweb.org … Figure 1 represents this process. 2Thanks to Stanford, for making the Stanford Classifier which is used as the maximum entropy classifier http://nlp. stanford. edu/software/classifier. shtml 279 Page 286. Figure 1: Architecture of JMaxAlign 3.1. … Cited by 1 Related articles All 7 versions

Final Project Stock Price Prediction Using News Articles Q Ma – 2008 – nlp.stanford.edu … The predicting power of textual information on financial markets. IEEE Intelligent Informatics Bulletin, 5(1):1–10, 2005. [4] Christopher Manning and Dan Klein. Stanford-classifier-2.0. http://nlp.stanford.edu/software/classifier.shtml. [5] Christopher Manning and Dan Klein. … Related articles All 2 versions

Catching metaphors M Gedigian, J Bryant, S Narayanan… – Proceedings of the Third …, 2006 – dl.acm.org … Steve Sinha and Srini Narayanan. 2005. Model-based answer selection. In Proceedings of the AAAI Work- shop on Inference for Textual Question Answering. Stanford Classifier. 2003. http://nlp.stanford.edu/software/classifier.shtml. 48 Cited by 58 Related articles All 12 versions

EDNII Error Detection Tool User Manual NLT Nguyen, Y Miyao – 2012 – edcw2012-open.googlecode.com … Selling point: ? Islam model is claimed to be able to correct multiple types of errors. External data and tools a. Stanford Classifier URL: http://nlp.stanford.edu/software/classifier.shtml Purpose: To train the multi-class classifier used in Classifier Model b. Stanford Parser … Related articles All 2 versions

Identifying protein-protein interaction sites using covering algorithm X Du, J Cheng, J Song – International journal of molecular sciences, 2009 – mdpi.com … method. LIBSVM is used as the SVM implementation with radial basis function as kernel and default C, ?. Stanford Classifier (ME) is used and can be download freely from http://www-nlp.stanford.edu/software/classifier.shtml. … Cited by 4 Related articles All 14 versions

A Hybrid Approach to Learn Description Logic based Biomedical Ontology from Texts? Y Ma, A Syamsiyah – ceur-ws.org … Multiclass classification is to answer whether a pair of two concept names has a relation, and if yes, which relation it is. This part can be achieved by the model learned from training data by Stanford Classifier. … 1 http://nlp.stanford.edu/software/classifier.shtml Page 4. … Related articles

Evaluation of Invalid Input Discrimination Using Bag-of-Words for Speech-Oriented Guidance System H Majima, R Torres, H Kawanami, S Hara… – Natural Interaction with …, 2014 – Springer … 4.2 [7] AM, LM Takemaru-model [3] Output 10-best candidates Morphological analyzer Chasen 2.3.3 [8] SVM Tool LIBSVM [5] Kernel function Radial basis function (RBF) Parameter C 10 -2,10 -1,…,104 ME Tool Stanford Classifier 2.1.3 … http://nlp.stanford.edu/software/classifier. … Related articles All 5 versions

Tri-Layer-Cluster Generation Model for Activity Prediction D Zhu, Y Fukazawa, J Ota – Web Intelligence (WI) and …, 2013 – ieeexplore.ieee.org … The Stanford Classifier was put forward as the control group, and the activity prediction accuracy demonstrates that the proposed model exhibits the superiority in multi-activity prediction. … As for quantitative evaluation, the Stanford Classifier [25] acted as a control group in our … Cited by 1 Related articles All 4 versions

A study of automatic email routing for an information technology help desk J Di Febo – 2013 – ideals.illinois.edu … and suffixes. Incorporating these features into a model can improve performance, as is seen for the Stanford Classifier in the classic 20 Newsgroups dataset [8] [9]. By carefully choosing … To this end, the Stanford Classifier has many features that make it well-suited for this task. … Related articles All 2 versions

UPC-CORE: What can machine translation evaluation metrics and Wikipedia do for estimating semantic textual similarity? LA Barrón Cedeño, L Màrquez Villodre, M Fuentes Fort… – 2013 – upcommons.upc.edu … Nor- malizing according to the entire dev-test dataset led to the best results 4We used the Stanford classifier; http://nlp. stanford.edu/software/classifier.shtml Table 2: Tuning process: parameter definition and feature selection. … Related articles All 2 versions

Sentiment analysis: Facebook status messages JK Ahkter, S Soria – Unpublished master’s thesis, Stanford, …, 2010 – www-nlp.stanford.edu … test sets. For this project, we used the Stanford Classifier v2.0 (the MaxEnt classifier), Stanford Tagger v3.0 (2010-05-26, English- only) (the POS tagger), and the Stanford Topic Modeling Toolbox v0.2.1 (the Labeled LDA). All … Cited by 13 Related articles All 3 versions

ISM @FIRE-2014: Named Entity Recognition Indian Languages S Dubey, B Goel, DK Prabhakar, S Pal – isical.ac.in … NER systems are extremely useful in many Natural Lan- guage Processing (NLP) applications such as question an- swering, machine translation, information extraction and so on. A Conditional Random Field (CRF) based Stanford Classifier has been used for classification. … Related articles

Semi-supervised SRL system with Bayesian inference A Lorenzo, C Cerisara – Computational Linguistics and Intelligent Text …, 2014 – Springer … these supervised models are used to produce a set of candidate semantic arcs on the unlabeled corpus. 1 We use the Stanford Classifier: http://nlp.stanford.edu/software/ classifier.shtml Page 4. 432 A. Lorenzo and C. Cerisara … Related articles All 6 versions

Evaluation of invalid input discrimination using BOW for speech-oriented guidance system H Majima, R Torres, H Kawanami, S Hara, T Matsui… – 2012 – library.naist.jp … this context, such features, for instance, could be the ut• terances’ G??1 likelihood, duration or SNR, which will be shown later.?’i are the parameters that need to be estirnated, which reflect the irnportance of /i (e, d) in the prediction For this work, Stanford Classifier [6] is … Related articles All 2 versions

Final Project: Sentiment analysis of news articles for financial signal prediction JJ Zhai, NN Cohen, A Atreya – dosen.narotama.ac.id … We developed Java data processing code and used the Stanford Classifier to quickly analyze financial news articles from The New York Times and predict sentiment in the articles. Two approaches were taken to produce sentiments for training and testing. … Related articles All 4 versions

Comparing methods for the syntactic simplification of sentences in information extraction RJ Evans – Literary and Linguistic Computing, 2011 – ALLC … This motivated the development of a hybrid classifier (HYBRID) that uses the MBL classifier when processing conjunctions and commas and uses the STANFORD classifier when processing adjacent comma–conjunction pairs. 3.2.4 Majority class baseline classifier. … Cited by 8 Related articles All 7 versions

Evaluation of the Delta TF-IDF Features for Sentiment Analysis AB Samoylov – Analysis of Images, Social Networks and Texts, 2014 – Springer … For example Stanford classifier 1 makes mistakes when dealing with the treatment of such phrases as: “I don’t love the film” and “I have not been disappointed by the film!”, classifying them as positive and negative. The NB classifier 2 gives the same result. … Related articles

Analyzing users’ narratives to understand experience with interactive products AN Tuch, R Trusell, K Hornbæk – … of the SIGCHI Conference on Human …, 2013 – dl.acm.org … We used a freely available implementation of such a classifier, the Stanford Classifier (http://nlp.stanford.edu/software/classifier.shtml, [14]). We used the classifier on default settings; the only advanced settings we used were SplitWordPairs and SplitWordShape. … Cited by 14 Related articles All 3 versions

Detecting ‘Request Alternatives’ User Dialog Acts from Dialog Context Y Ma, E Fosler-Lussier – uni-ulm.de … reqalts). We train a maximum entropy classifier (in particular, we use the Stanford Classifier [3]) as it can discern Proceedings of 5th International Workshop on Spoken Dialog Systems Napa, January 17-20, 2014 85 Page 3. … Related articles All 2 versions

Determining latency for on-line dialog act classification S Germesin, T Becker, P Poller – Poster Session for the 5th International …, 2008 – ami.dfki.de … Hence, we limited the amount of updatings per new classified segments. 2 http://http://nlp.stanford. edu/software/classifier.shtml 3 A livelock is similar to a deadlock, except that the states of the processes involved in the livelock constantly change but none is progressing. Page 4. … Cited by 8 Related articles All 3 versions

FIRE @ISM-2013 Transliterated Search Task DK Prabhakar, S Pal – irsi.res.in … 17:1-46 5. Dale, R.: Language Technology. Slides of HCSNet Summer School Course. Sydney (2007) 6. Stanford Classifier v3.2.0 – 2013-06-19 classification tool from Stanford University Page 16. 04/12/13 16 THANK YOU Related articles All 4 versions

FeatureForge: A Novel Tool for Visually Supported Feature Engineering and Corpus Revision. F Heimerl, C Jochim, S Koch, T Ertl – COLING (Posters), 2012 – Citeseer … FeatureForge currently integrates the feature definition language of the ColumnDataClassifier which is part of the Stanford Classifier suite (Manning and Klein, 2003) and makes them available to users, with the restriction of only allowing Boolean feature definitions, resulting in … Cited by 1 Related articles All 6 versions

Hey #311, Come Clean My Street!: A Spatio-temporal Sentiment Analysis of Twitter Data and 311 Civil Complaints R Eshleman, H Yang – Big Data and Cloud Computing ( …, 2014 – ieeexplore.ieee.org … 65.0 Stanford NLP0 N/A N/A N/A N/A 69.900 Table 2: Summary of Sentiment Classifier Performance. *The Stanford Classifier is not part of the WEKA suite. **The Stanford Classifier was not trained with our training data. B. Inter …

Analysing customers sentiments: An approach to opinion mining and classification of online hotel J Sixto, A Almeida, D López-de-Ipina – morelab.deusto.es … 1 http://wordnet.princeton.edu/ 2 http://nlp.stanford.edu/software/classifier.shtml 3 http://www.textfixer.com/resources/common-english-words.txt Page 3. Figure 1. Structure of proposed framework by Q-Wordnet. If a noun has several adjectives, these are evaluated … Related articles All 2 versions

Analysing Customers Sentiments: An Approach to Opinion Mining and Classification of Online Hotel Reviews J Sixto, A Almeida, D López-de-Ipiña – Natural Language Processing and …, 2013 – Springer … Using part-of-speech (POS) tagger information, we are able to relate an ad- jective with the noun it refers to, generating bigrams with a polarity assigned 1 http://wordnet.princeton.edu/ 2 http://nlp.stanford.edu/software/classifier.shtml 3 http://www.textfixer.com/resources/common … Related articles All 5 versions

Question Classification G Aggarwal, R Bhayani, T Tenneti – 2009 – cs229.stanford.edu … We simply use add-1 smoothing in our project and it works well. 4.2.2 MaxEnt We made use of Stanford classifier to code the Max- ent section. The idea behind MaxEnt classifiers is that we should prefer the most uniform models that satisfy any given constraint. … Related articles All 3 versions

Learning ensembles of structured prediction rules C Cortes, V Kuznetsov, M Mohri – Proceedings of ACL, 2014 – cs.nyu.edu Page 1. Learning Ensembles of Structured Prediction Rules Corinna Cortes Google Research 111 8th Avenue, New York, NY 10011 corinna@google.com Vitaly Kuznetsov Courant Institute 251 Mercer Street, New York, NY 10012 vitaly@cims.nyu.edu … Cited by 3 Related articles All 6 versions

Question classification for email R Cotterill – Proceedings of the Ninth International Conference on …, 2011 – dl.acm.org … supplied with QANUS v26Jan2010. The question classification component of QA-SYS is an instance of the Stanford classifier, a supervised learning module trained on a dataset of information retrieval questions. Ng & Kan do … Related articles All 8 versions

DCU and ISI @INEX 2010: Adhoc and data-centric tracks D Ganguly, J Leveling, GJF Jones… – … Evaluation of Focused …, 2011 – Springer … machine’- processed set), is similar, but negation was now automatically identified using a Maximum Entropy Classifier (Stanford Classifier) 3, and … Q and P are statistically significant, corroborating the fact that removal of negation 3 http://nlp.stanford.edu/software/classifier.shtml … Cited by 1 Related articles All 15 versions

Final Project Information Extraction from Housing Advertisements D Murray, J Herbach, R Jain – 2009 – wiki.eecs.yorku.ca … 3 Jun. 2009 <http://nlp.stanford.edu/software/index.shtml>. 2. “The Stanford Classifier”. The Stanford NLP Group. 2009. Stanford University. 3 Jun. 2009 <http://nlp.stanford.edu/software/ classifier.shtml>. 3. “The Stanford POS Tagger”. The Stanford NLP Group. 2009. … Related articles All 4 versions

Recognising sentence similarity using similitude and dissimilarity features S Sangeetha, M Arock – International Journal of Advanced Intelligence …, 2012 – Inderscience … 4.2.6 Classifier We have employed the Maximum entropy classifier (Manning and Klein, 2003; http://www.nlp.stanford.edu/software/classifier.shtml) for classification task. This ME classifier learns from the training set and classifies test set into paraphrases or non- paraphrases. … Cited by 1 Related articles All 3 versions

“My Curiosity was Satisfied, but not in a Good Way”: Predicting User Ratings for Online Recipes C Liu, C Guo, D Dakota, S Rajagopalan, W Li… – SocialNLP 2014, 2014 – aclweb.org … To aggregate review-level ratings into a recipe-level prediction, we experimented with both the max- imum entropy classifier in the implementation of the Stanford Classifier (Manning and Klein, 2003) and the SVM multi-class classifier. … Related articles All 3 versions

Sentiment Classification in Chinese Microblogs: Lexicon-based and Learning-based Approaches B Yuan, Y Liu, H Li, TTT PHAN, G Kausar… – … and Research (IPEDR), 2013 – ipedr.com … We design a dataset consisting of 4 Domain Specific Datasets (DSD) and 1 Random Dataset (RD). Those datasets have been labelled with polarity for 4 weeks in a row by 2 annotators, who are both graduate 4 http://nlp.stanford.edu/software/classifier.shtml 3 Page 4. … Cited by 3 Related articles All 4 versions

A meta-top-down method for large-scale hierarchical classification XL Wang, H Zhao, BL Lu – Knowledge and Data Engineering, …, 2014 – ieeexplore.ieee.org Page 1. A Meta-Top-Down Method for Large-Scale Hierarchical Classification Xiao-Lin Wang, Hai Zhao, and Bao-Liang Lu, Senior Member, IEEE Abstract—Recent large-scale hierarchical classification tasks typically have tens … Cited by 2 Related articles All 10 versions

Question Classification for the Icelandic Language: First Steps ÓP Geirsson – geirsson.com … For this part, we only used the Gettu Betur corpus. 3.1 Training model Maximum entropy models have shown to work well on text classification (Berger et al., 1996). In the fol- lowing, the Stanford classifier (Manning and Klein, 2003) is used. … Related articles

RU:-) or:-(? character-vs. word-gram feature selection for sentiment classification of OSN corpora B Blamey, T Crick, G Oatley – … and Development in Intelligent Systems XXIX, 2012 – Springer … Looking more closely at our data, we can see that character n-grams consistently beat word unigrams, which is understandable, as 8 characters will often be enough 3 http://nlp.stanford. edu/software/classifier.shtml 4 http://svmlight.joachims.org/ 5 http://alias-i.com/lingpipe/ 210 … Cited by 3 Related articles All 7 versions

Automatic prediction of text aesthetics and interestingness D Ganguly, J Leveling, GJF Jones – 2014 – doras.dcu.ie … 8http://mallet.cs.umass.edu/ 9http://nlp.stanford.edu/software/tagger.shtml 10http://sentiwordnet. isti.cnr.it/code/SentiWordNetDemoCode.java 11http://www.linguatools.de/disco/disco_en.html 12http://nlp.stanford.edu/software/classifier.shtml 13http://www … by the Stanford classifier. … Related articles All 6 versions

Learning formal definitions for Snomed CT from text Y Ma, F Distel – Artificial Intelligence in Medicine, 2013 – Springer … PhD thesis, Universität Karlsruhe (2009) 8. Wächter, T., Fabian, G., Schroeder, M.: Dog4dag: Semi-automated ontology genera- tion in OBO-Edit and Protégé. In: Proceedings of SWAT4LS 2011, pp. 119–120 (2011) 3 http://nlp.stanford.edu/software/classifier.shtml Cited by 7 Related articles All 8 versions

Cross-linguistic annotation of narrativity for English/French verb tense disambiguation C Grisot, T Meyer – 9th Edition of the Language Resources and …, 2014 – infoscience.epfl.ch … The 435 correctly annotated instances of narrativity (257 narrative, 178 non-narrative), after resolving the disagree- ments as described in Section 4, have been used entirely for training a Maximum Entropy (MaxEnt) classifier with the Stanford Classifier package (Manning and … Cited by 1 Related articles All 3 versions

Classification ensemble to improve medical Named Entity Recognition S Keretna, CP Lim, D Creighton… – Systems, Man and …, 2014 – ieeexplore.ieee.org … A. Development tools The ME classifier used in this experiment is Stanford classifier V3.3.1 [33]. It is an open sources package under GNU general public license. For the CRF classifier, Stanford NER V3.3.1 is used [34], which is published under full GPL license. … Related articles

Lazy sparse stochastic gradient descent for regularized multinomial logistic regression B Carpenter – Alias-i, Inc., Tech. Rep, 2008 – Citeseer Page 1. Lazy Sparse Stochastic Gradient Descent for Regularized Mutlinomial Logistic Regression Bob Carpenter Alias-i, Inc. carp@alias-i.com Abstract Stochastic gradient descent efficiently estimates maximum likelihood logistic regression coefficients from sparse input data. … Cited by 24 Related articles

Opinion Mining in Conversational Content within Web Discussions and Commentaries K Machová, L Marhefka – … , Reliability, and Security in Information Systems …, 2013 – Springer … Detection in Short Informal Text. Journal of the American Society for Information Science and Technology 61(12), 2544–2558 (2010) 13. Stanford Classifier. Stanford University, http://nlp.stanford.edu/software/classifier.shtml Related articles All 4 versions

Improving Persian POS tagging using the maximum entropy model AA Kardan, MB Imani – Intelligent Systems (ICIS), 2014 Iranian …, 2014 – ieeexplore.ieee.org … The tag set distribution is presented in Fig. 2[7]. Figure 2. Tag Distribution for BijanKhan dataset (The tags which are picked for “ETC” group are the ones whose number of occurrences is below 5000 times in the corpus)[11] 1http://nlp.stanford.edu/software/classifier.shtml … Related articles

RelANE: Discovering Relations between Arabic Named Entities I Boujelben, S Jamoussi, AB Hamadou – Text, Speech and Dialogue, 2014 – Springer … According to the empirical results illustrated in this table, the used algorithms ob- tained very competitive scores. The highest performance of our system is accomplished 5 http://www.cs. waikato.ac.nz/ml/weka/ 6 http://nlp.stanford.edu/software/classifier.shtml Page 6. … Related articles All 5 versions

Choosing the right translation: A syntactically informed classification approach S Zwarts, M Dras – Proceedings of the 22nd International Conference on …, 2008 – dl.acm.org … Our idea for estimating the wrongness of a parse, or the complexity of a parse that might lead to incorrect reordering rule application, is to use ‘side-effect’ informa- 1http://svmlight.joachims. org 2http://nlp.stanford.edu/software/classifier.shtml 1156 Page 5. … Cited by 13 Related articles All 12 versions

Report of Working Group on Literature, Lexicon, Diachrony L Auvil, D Bamman, C Brown, G Crane… – … –Bridging the Gap … – drops.dagstuhl.de … These include: the determination of error rate and causes of error in the application of the Stanford Classifier to the identification of group names in the English texts of the Perseus corpus of ancient authors; the refinement of the classifier to deal with authors of different genres … Related articles All 2 versions

Intelligent News Aggregator for German with Sentiment Analysis D Ploch – Smart Information Systems, 2015 – Springer … laneous. Since the Stanford classifier sometimes misses some named entities we decided to augment the list of named entities returned by the Stanford classifier by the named entities identified by the part-of-speech tagger described above. … Related articles All 3 versions

QANUS: An Open-source Question-Answering Platform JP Ng, MY Kan – arXiv preprint arXiv:1501.00311, 2015 – arxiv.org … We built the classifier used by training the Stanford Classifier (Manning and Klein, 2003) on the data described in Li and Roth (2002). The classification assigned to each question is stored and passed on to the answer re- trieval stage. Answer Retrieval. … Cited by 3 Related articles All 11 versions

Visualization of live search O Nilsson – 2013 – diva-portal.org Page 1. Institutionen för datavetenskap Department of Computer and Information Science Final thesis Visualization of live search by Olof Nilsson LIU-IDA/LITH-EX-A–13/002–SE 2013-12-09 Linköpings universitet SE-581 83 Linköping, Sweden … Related articles

Detecting narrativity to improve English to French translation of simple past verbs T Meyer, C Grisot… – Proceedings of the 1st …, 2013 – infoscience.epfl.ch … class agreement. 3.2 Features for Narrativity The manually annotated instances were used for training and testing a Maximum Entropy classi- fier using the Stanford Classifier package (Man- ning and Klein, 2003). We extracted … Cited by 4 Related articles All 11 versions

UM-Checker: A Hybrid System for English Grammatical Error Cor-rection J Xing, L Wang, DF Wong, LS Chao, X Zeng – CoNLL-2013, 2013 – cms.kdis.edu.cn Page 44. Proceedings of the Seventeenth Conference on Computational Natural Language Learning: Shared Task, pages 34–42, Sofia, Bulgaria, August 8-9 2013. cO2013 Association for Computational Linguistics UM-Checker … Cited by 9 Related articles All 10 versions

Learning explicit and implicit Arabic discourse relations I Keskes, FB Zitoune, LH Belguith – … of King Saud University-Computer and …, 2014 – Elsevier … Instead, EDUs are automatically identified and then manually corrected if necessary. The segmentation of our corpus was performed by a multi-class supervised learning approach using the Stanford classifier that is based on the Maximum Entropy model (Ratnaparkhi, 1997). … Related articles

Subcategorisation Acquisition from Raw Text for a Free Word-Order W Roberts, M Egg, V Kordoni – EACL 2014, 2014 – aclweb.org Page 324. Proceedings of the 14th Conference of the European Chapter of the Association for Computational Linguistics, pages 298–307, Gothenburg, Sweden, April 26-30 2014. cO2014 Association for Computational Linguistics … Related articles All 4 versions

Pro, Con, and Affinity Tagging of Product Reviews T Sullivan – www-nlp.stanford.edu … The maximum entropy classifier allows us to eas- ily add many features to constrain the current data instance while leaving the rest of the probabilities pleasantly uniform (equally likely). We used the Stanford Classifier [10] as our out-of-the-box maximum entropy clas- sifier. … Cited by 1 Related articles All 3 versions

Merging Lexicons for Higher Precision Subcategorization Frame Acquisition L Rimell, T Poibeau, A Korhonen – LREC 2012 Workshop on …, 2012 – lrec.elra.info … Not all of these sen- tences were classified in each pipeline, either due to parser errors or to the GRs failing to match the rules for any SCF. On average, the RASP classifier classified 6500 sentences per verb, the Stanford classifier 5594, and the combined classifier only 1922. … Cited by 1 Related articles All 9 versions

SRIUBC: simple similarity features for semantic textual similarity E Yeh, E Agirre – Proceedings of the First Joint Conference on Lexical …, 2012 – dl.acm.org … We used the Stanford Classifier’s (Manning and Klein, 2003) multinomial logistic regression as our dataset predictor, using the feature vectors from System 2. Five-fold cross validation over the training data showed the dataset predictor to have an overall ac- curacy of 91.75%. … Cited by 1 Related articles All 12 versions

Veridicality and utterance understanding M de Marneffe, CD Manning… – … Computing (ICSC), 2011 …, 2011 – ieeexplore.ieee.org … In all our experiments, we use the Stanford Classifier [19] with Gaussian prior N(0,1). For clas- sification tasks, the dominant tradition within computational linguistics has been to adjudicate differing human judgments and to assign a single class for each item in the training data. … Cited by 3 Related articles All 14 versions

Tuning as ranking M Hopkins, J May – Proceedings of the Conference on Empirical Methods …, 2011 – dl.acm.org Page 1. Proceedings of the 2011 Conference on Empirical Methods in Natural Language Processing, pages 1352–1362, Edinburgh, Scotland, UK, July 27–31, 2011. c 2011 Association for Computational Linguistics Tuning as Ranking … Cited by 106 Related articles All 11 versions

[BOOK] Verb Polarity Frames: a New Resource and Its Application in Target-specific Polarity Classification M Klenner, M Amsler, N Hollenstein, G Faaß – 2014 – opus.bsz-bw.de Page 115. Verb Polarity Frames: a New Resource and its Application in Target-specific Polarity Classification Manfred Klenner Computational Linguistics University of Zurich Switzerland klenner@ cl. uzh. ch Michael Amsler … Related articles

Social media sentiment analysis and topic detection for Singapore English YL Phua – 2013 – DTIC Document … more challenging to annotate. Some of the NLP work on Facebook includes [19] which used Stanford Classifier, Stanford Tagger and Stanford Topic Modeling Toolbox for sentiment analysis and [20] that performed real time opinion … Cited by 1 Related articles All 2 versions

A cascaded classification approach to semantic head recognition L Michelbacher, A Kothari, M Forst, C Lioma… – Proceedings of the …, 2011 – dl.acm.org … fold. Those fea- tures were amt, amf and amscp. The main experiment combines these three se- 4http://nlp.stanford.edu/software/ classifier.shtml 798 Page 7. lected AM features with all possible subsets of con- text features. We … Cited by 4 Related articles All 5 versions

Automatic Detection of Point of View Differences in Wikipedia. K Al Khatib, H Schütze, C Kantner – COLING, 2012 – aclweb.org … 40 Page 9. For English, BOW and n-gram features are directly computed from text (as tokenized by the Stanford classifier) without any further linguistic preprocessing like lemmatization. For Arabic, we investigate a number of different options for linguistic preprocessing. … Cited by 2 Related articles All 5 versions

A survey of Arabic question answering: Challenges, tasks, approaches, tools, and future trends A Ezzeldin, M Shaheen – Proceedings of The 13th International …, 2012 – researchgate.net … Stanford Word Segmenter: This is a CRF-based word segmenter in Java which also supports Arabic and Chinese. – Stanford Classifier: This is a machine learning classifier for text categorization, a maximum entropy and multi-class logistic regression model. … Cited by 3 Related articles All 2 versions

Did it happen? The pragmatic complexity of veridicality assessment MC de Marneffe, CD Manning, C Potts – Computational linguistics, 2012 – MIT Press Cited by 22 Related articles All 13 versions

Predicting Stance in Ideological Debate with Rich Linguistic Knowledge KSHV NG – 24th International Conference on Computational …, 2012 – aclweb.org Page 465. Proceedings of COLING 2012: Posters, pages 451–460, COLING 2012, Mumbai, December 2012. Predicting Stance in Ideological Debate with Rich Linguistic Knowledge Kazi Saidul HASAN Vincent NG Human Language … Related articles All 8 versions

Why are You Taking this Stance? Identifying and Classifying Reasons in Ideological Debates KS Hasan, V Ng – aclweb.org … scores for each domain. Let us begin by discussing the RC results. First, P1 and P2 significantly beat the Baseline on all 4http://nlp.stanford.edu/software/ classifier.shtml 5http://sourceforge.net/projects/ lpsolve/ four domains by an … Related articles All 6 versions

Reusable research? a case study in named entity recognition M Van Erp, L Van der Meij – 2013 – wordpress.let.vupr.nl … In particular for the ‘smaller’ data elements (ie, those making up a smaller proportion of the dataset) the Stanford classifier outperforms the [1] approach significantly, probably because it is unaware of the fact that this data belongs to a different element. … Cited by 1 Related articles

Arabic Question Answering: Systems, Resources, Tools, and Future Trends M Shaheen, AM Ezzeldin – Arabian Journal for Science and Engineering, 2014 – Springer Page 1. Arab J Sci Eng DOI 10.1007/s13369-014-1062-2 REVIEW ARTICLE – COMPUTER ENGINEERING AND COMPUTER SCIENCE Arabic Question Answering: Systems, Resources, Tools, and Future Trends Mohamed Shaheen • Ahmed Magdy Ezzeldin … Related articles All 2 versions

CS224N/Ling237 Final Project E Yeh – www-nlp.stanford.edu … Manning, Christopher and Dan Klein. 2003. Optimization, Maxent Models, and Conditional Estimation without Magic. Tutorial at HLT-NAACL 2003 and ACL 2003, http://www- nlp.stanford.edu/software/classifier.shtml. Manning, Christopher and Dan Klein. 2005. … Related articles All 3 versions

Ensemble methods for structured prediction C Cortes, V Kuznetsov, M Mohri – Proceedings of the …, 2014 – machinelearning.wustl.edu Page 1. Ensemble Methods for Structured Prediction Corinna Cortes CORINNA@ GOOGLE.COM Google Research, 111 8th Avenue, New York, NY 10011 Vitaly Kuznetsov VITALY@CIMS.NYU.EDU Courant Institute of Mathematical … Cited by 3 Related articles All 5 versions

Review Selection Using Micro-Reviews T Nguyen, H Lauw, P Tsaparas – 2014 – ieeexplore.ieee.org Page 1. 1041-4347 (c) 2013 IEEE. Personal use is permitted, but republication/ redistribution requires IEEE permission. See http://www.ieee.org/ publications_standards/publications/rights/index.html for more information. This … Related articles All 4 versions

Using micro-reviews to select an efficient set of reviews TS Nguyen, HW Lauw, P Tsaparas – Proceedings of the 22nd ACM …, 2013 – dl.acm.org … Because topic modeling is probabilistic, we average the semantic similarity over ten runs. To determine the sentiment polarity of each sentence and tip, we train a sentiment classifier using the Stanford Classifier toolkit [15] with textual features (word and letter n-grams). … Cited by 2 Related articles All 5 versions

Ensemble Methods for Structured Prediction V Kuznetsov, M Mohri – cs.nyu.edu Page 1. Ensemble Methods for Structured Prediction Corinna Cortes CORINNA@ GOOGLE.COM Google Research, 111 8th Avenue, New York, NY 10011 Vitaly Kuznetsov VITALY@CIMS.NYU.EDU Courant Institute of Mathematical … Related articles

Applying Conditional Random Fields into Bioinformatics W Liu – 2007 – rp-www.cs.usyd.edu.au Page 1. Applying Conditional Random Fields into Bioinformatics Wei Liu SID: 307035077 Master of Science by Research School of Information Technologies Supervisor: Dr Sanjay Chawla The University of Sydney 5 November 2007 Page 2. Abstract … Related articles All 2 versions

Managing Information Extraction in a Mashup Environment By: Vasiliki P. Prokopi Advisor: Professor Minos Garofalakis Co-advisor: Professor Stavros Christodoulakis Co-advisor: Professor Antonios Deligiannakis TECHNICAL UNIVERSITY OF CRETE … Related articles All 2 versions

Using negative information in search S Palchowdhury, S Pal, M Mitra – Emerging Applications of …, 2011 – ieeexplore.ieee.org … LNCS, vol. 6203, 2010. [10] A. Rafferty and C. Manning, “Stanford classifier,” http://nlp.stanford. edu/software/classifier.shtml. [11] G. Salton, Ed., The SMART Retrieval System—Experiments in Automatic Document Retrieval. Prentice Hall Inc., Englewood Cliffs, NJ, 1971. … Cited by 1 Related articles All 4 versions

Spoken dialogue systems K Jokinen, M McTear – Synthesis Lectures on Human …, 2009 – morganclaypool.com Page 1. Spoken Dialogue Systems Page 2. Page 3. iii Synthesis Lectures on Human Language Technologies Editor Graeme Hirst, University of Toronto Synthesis Lectures on Human Language Technologies publishes monographs … Cited by 44 Related articles All 9 versions

An evaluation of classification models for question topic categorization B Qu, G Cong, C Li, A Sun… – Journal of the American …, 2012 – Wiley Online Library Skip to Main Content. Wiley Online Library. Log in / Register. Log In E-Mail Address Password Forgotten Password? Remember Me. … Cited by 14 Related articles All 8 versions

Final Project: A Semantic, Supervised Classification Approach to Restaurant Reviews P Vantimitta – nlp.stanford.edu Page 1. [1] CS 224n Final Project: A Semantic, Supervised Classification Approach to Restaurant Reviews Pavani Vantimitta pavani@stanford.edu Abstract The rapid growth of E-commerce has made Customer reviews an indispensable … Related articles All 4 versions

Preference Extraction From Negotiation Dialogues. A Cadilhac, N Asher, F Benamara, V Popescu, M Seck – ECAI, 2012 – books.google.com Page 241. Preference Extraction From Negotiation Dialogues Anais Cadilhac, Nicholas Asher, Farah Benamara, Vladimir Popescu and Mohamadou Seck1 Abstract. This paper presents an NLP-based approach to extract- ing preferences from negotiation dialogues. … Cited by 1 Related articles All 11 versions

A Technical Report on Approaches for Information Extraction and Sentiment Mining S Morsy, N Ayman, I Elnabarawy, M Shalaby, A Hamed… – cse.aucegypt.edu Page 1. A Technical Report on Approaches for Information Extraction and Sentiment Mining Project Title Sentiment Analysis and Opinion Mining of the Arabic Web (Digital Content) Selected ITAC Program Advanced Research Project (ARP) Academic and ICT Industry Partners … Related articles

Barrier and Syntactic Features for Information Retrieval A Alicante – 2013 – fedoa.unina.it Page 1. Tesi di Dottorato Anita Alicante March 2013 Dottorato in Scienze Computazionali e Informatiche XXV Ciclo Universitá degli studi di Napoli “Federico II” Barrier and Syntactic Features for Information Retrieval Page 2. Page 3. Barrier and Syntactic Features … Related articles All 3 versions

[BOOK] An Analysis of the Use of Syntax in Statistical Machine Translation S Zwarts – 2008 – comp.mq.edu.au Page 1. An Analysis of the Use of Syntax in Statistical Machine Translation Simon Zwarts Master of Science (M.Sc) This thesis is presented for the degree of Doctor of Philosophy at the Department of Computing Division ICS Macquarie University October 2008 Page 2. ii Page 3. … Related articles