Preview

Automatic Emotion Recognition from Speech Using Reduced Feature Set & Different Classifiers

Good Essays
Open Document
Open Document
1309 Words
Grammar
Grammar
Plagiarism
Plagiarism
Writing
Writing
Score
Score
Automatic Emotion Recognition from Speech Using Reduced Feature Set & Different Classifiers
Automatic Emotion Recognition from Speech using Reduced Feature Set & Different Classifiers

U09CO202 U09CO207

U09CO206 U09CO240

Outline
● Project Preliminary : A quick recap ● Running the SVM classifier - Weka ● Improvising the baseline model ○ Principal Component Analysis ○ Feature Subset Selection ● Comparison of different models ● Building a local database ● Next Steps...

Outline
● Project Preliminary : A quick recap ● Running the SVM classifier - Weka ● Improvising the baseline model ○ Principal Component Analysis ○ Feature Subset Selection ● Comparison of different models ● Building a local database ● Next Steps...

So far ...
● Used the Berlin speech emotion database Emo-DB: http://www.expressive-speech.net/emodb/ ● Feature extraction : openSMILE ● Trained multi-class SMO classifer in Weka ○ Accuracy : 82.4%

Workflow

Outline
● Project Preliminary : A quick recap ● Running the SVM classifier - Weka ● Improvising the baseline model ○ Principal Component Analysis ○ Feature Subset Selection ● Comparison of different models ● Building a local database ● Next Steps...

Weka SMO parameters
We used the foll. configuration in the Classify window of the Weka Explorer : weka . classifiers . functions . SMO −C 1 . 0 −L 0 . 0 0 1 −P 1 . 0 E−12 −N 0 −V −10 −W 1 −K ” weka .classifiers . functions . supportVector . PolyKernel −C 250007 −E 1 . 0 ”

weka . classifiers . functions . SMO −C 1 . 0 −L 0 . 0 0 1 −P 1 . 0 E−12 −N 0 −V −10 −W 1 −K ” weka .classifiers . functions . supportVector . PolyKernel −C 250007 −E 1 . 0 ”

● weka . classifiers . functions . SMO : We are using the SMO class of the Weka jar library as our classifier ● -C 1.0 : Cost parameter for the soft margin SVM classifier. Balances between misclassifications & rigid boundaries. ● -L 0.001 : Tolerance parameter set for SMO convergence. Default ● -P 1.0e-12: Epsilon that determines number of SVs to choose. Larger P => smaller % of SVs

weka . classifiers . functions .

You May Also Find These Documents Helpful

  • Satisfactory Essays

    Pt1420 Unit 1 Assignment

    • 303 Words
    • 2 Pages

    IBM Multimedia Analysis and Retrieval System [8]. The service enabled users to train new classifiers in December 2015.…

    • 303 Words
    • 2 Pages
    Satisfactory Essays
  • Powerful Essays

    Solutions Chapter 7

    • 7531 Words
    • 30 Pages

    Objective Topic Edition Edition 31 LO 2 Gain recognition and basis computation Unchanged 31…

    • 7531 Words
    • 30 Pages
    Powerful Essays
  • Good Essays

    The Department of Justice uses two types of programs to measure crime in the United States. The two types of programs are the Uniform Crime Reporting Program (URC) and National Crime Victimization Survey (NCVS). Which both have valuable details about the crime problems that are going on in the world.Since both programs are used to do different things and methods, both do focus on different crimes and there aspect. The data that they produce together is used for comprehensive crime problems that are going on in the world.…

    • 1354 Words
    • 4 Pages
    Good Essays
  • Powerful Essays

    "The students, united, will never be defeated!". This was the rallying call of students at San Francisco State University trying to save their Asian America Studies classes in 2008. 40 years after the first Third World Liberation Front protested and fought against discrimination and for their rights to have ethnic studies classes, the fight still continues. Students fought to have classes from the perspective of ethnic peoples, and not the Eurocentric point of view. In doing so, the struggle for ethnic studies from students and the community challenged and broke the status quo and construct of "race" in a Eurocentric America.…

    • 1973 Words
    • 8 Pages
    Powerful Essays
  • Good Essays

    Accuracy Assessment Paper

    • 344 Words
    • 2 Pages

    Accuracy assessment is an important final step of the classification process. The goal is to…

    • 344 Words
    • 2 Pages
    Good Essays
  • Best Essays

    It Essay - Data Mining

    • 1998 Words
    • 8 Pages

    He, J. (2009). Advances in Data Mining: History and Future. Third International Symposium on Intelligent . Retrieved November 1, 2012, from http://ieeexplore.ieee.org.ezproxy.lib.ryerson.ca/stamp/stamp.jsp?tp=&arnumber=5370232&tag=1…

    • 1998 Words
    • 8 Pages
    Best Essays
  • Satisfactory Essays

    In order to explain the use of various algorithms in this study, the algorithms will be discussed in this research. Naïve Bayes and Apriori will be used against the Stylometry data set. IBk will be used against the Keystroke Capture and Mouse Movement data sets. J48 will be used with the Mushroom Database. The choices of these techniques and their implementation will be discussed in detail in the methodologies section. According to Witten and Frank in Data Mining, the Naïve Bayes method is, “based on Bayes’srule and ‘Naïvely’ assumed independence — it is only valid to multiply probabilities when the According to Witten and Frank in Data Mining, the Naïve Bayes method is, “based on Bayes’s rule and ‘Naïvely’ assumed independence — it is only valid to multiply probabilities when the events are independent. The assumption that attributes are independent in real life certainly is simplistic one events are independent. The assumption that attributes are independent in real life certainly is a simplistic one.…

    • 494 Words
    • 2 Pages
    Satisfactory Essays
  • Better Essays

    We use many models to make classification and prediction. The three models are multiple linear regression, classification tree and neural network.…

    • 1929 Words
    • 8 Pages
    Better Essays
  • Powerful Essays

    Data Mining

    • 1921 Words
    • 8 Pages

    Patterson, L. (2010, APR 27). The nine most common data mining techniques used in predictive…

    • 1921 Words
    • 8 Pages
    Powerful Essays
  • Powerful Essays

    data sources on the topic of interest, the evaluation of this work in relation to the problem, and the documentation of the work.…

    • 1472 Words
    • 5 Pages
    Powerful Essays
  • Powerful Essays

    Abstract Support vector machines (SVMs) are promising methods for the prediction of ÿnancial timeseries because they use a risk function consisting of the empirical error and a regularized term which is derived from the structural risk minimization principle. This study applies SVM to predicting the stock price index. In addition, this study examines the feasibility of applying SVM in ÿnancial forecasting by comparing it with back-propagation neural networks and case-based reasoning. The experimental results show that SVM provides a promising alternative to stock market prediction. c 2003 Elsevier B.V. All rights reserved.…

    • 4964 Words
    • 20 Pages
    Powerful Essays
  • Powerful Essays

    Raiko, T., Ilin, A., & Karhunen, J. (2007). Principal component analysis for large scale problems with…

    • 4443 Words
    • 18 Pages
    Powerful Essays
  • Good Essays

    20 Kohavi, R., and John, G. H., 1997, “Wrappers for Feature Subset Selection,” Artif. Intell., 97 1–2 , pp. 273–324. 21 Espinosa, J., Vandewalle, J., and Wertz, V., 2005, Fuzzy Logic, Identification and Predictive Control, Springer-Verlag, London, UK. 22 http://en.wikipedia.org/wiki/Curse_of_dimensionality. 23 Bishop, C. M., 1995, Neural Networks for Pattern Recognition, Oxford University, New York. 24 Seidel, P., Seidel, A., and Herbarth, O., 2007, “Multilayer Perceptron Tumor Diagnosis Based on Chromatography Analysis of Urinary Nucleoside,” Neural Networks, 20 5 , pp. 646–651. 25 Smola, A. J., and Schoelkopf, B., 2004, “A Tutorial on Support Vector Regression,” Stat. Comput., 14 3 , pp. 199–222. 26 Cristianini, N., and Shawe-Taylor, J., 2000, An Introduction to Support Vector Machines and Other Kernel-Based Learning Methods, Cambridge University, New York, p. 189. 27 Prasad, A. M., Iverson, L. R., and Liaw, A., 2006, “Newer Classification and Regression Tree Techniques: Bagging and Random Forests for Ecological Prediction,” Ecosystems, 9 2 , pp. 181–189. 28 Breiman, L., 2001, “Random Forest,” Mach. Learn., 45 1 , pp. 5–32. 29 Breiman, L., Friedman, J., Olshen, R. A., and Stone, C. J., 1984, Classification and Regression Trees, Wadsworth International, Monterey, CA. 30 Wang, Y., and Witten, I. H., 2002, “Modeling for Optimal Probability Prediction,” Proceedings of the 19th International Conference in Machine Learning, Sydney, Australia, pp. 650–657.…

    • 6353 Words
    • 26 Pages
    Good Essays
  • Powerful Essays

    Personality Test

    • 3258 Words
    • 14 Pages

    Harris, Todd, Dr. "Science of the Predictive Index." PI Worldwide. N.p., 1 Jan. 2011. Web. 12 Feb. 2011. <http://www.piworldwide.com/Products/Predictive-Index-System/Science-of-the-Predictive-Index.aspx>.…

    • 3258 Words
    • 14 Pages
    Powerful Essays
  • Better Essays

    Solution Exercise

    • 682 Words
    • 3 Pages

    Homework #2 of IM5803701 – Data Mining and Applications, Spring 2012 Classification Data Mining Experiment, 100% Due date is on 5/1 9:30 AM by submitting on blackboard Guideline This homework is an “individual” data mining experiment. Plagiary is definitely not allowed. If any classmate or other person helps you on doing this homework, you need to specify who and which portion you got help from. You credit will be given to the helper (it is fair, right). The helper should also mention who get your assistant on this homework. Zero point will be given if your homework is found to be the same with others without any mention. You are required to use computer language (C, C++, or Java) or computer software (matlab or Weka) to do the data mining experiment and analysis. Other software or language is allowed based on the approval of the instructor. You need to specify which software or program you are using for this homework. If you use other person’s program or any program downloaded from internet, you need to address where you get it, and who is the author. If you decide to write your own program, please submit your source code. Extra credit will be given if you write your own program on any portion of this homework. No matter which kind of method you choose for this homework, you need to be careful on adjusting the parameter, if there is. Please do an experiment on how to obtain the better parameters and write down you analysis on this homework. You need to submit your homework written by MS Word through blackboard system. The homework should not longer than 10-page limit (source code should put in the appendix). No late homework is allowed! I. Congressional Voting Records (50%) http://archive.ics.uci.edu/ml/datasets/Congressional+Voting+Records Go to the UCI Machine Learning Repository to download the “Congressional Voting Records Data Set” or download house-votes-84.csv file from blackboard. Then, chose at least two different classification methods (decision tree,…

    • 682 Words
    • 3 Pages
    Better Essays