1. Emotion Detection in Email Customer Care Narendra Gupta, Mazin Gilbert, and Giuseppe Di Fabbrizio AT&T Labs - Research, Inc. 2010 ACL

2. Outline  [1.Introduction]  [2.Emotion detection in emails] 2.1 Classifier 2.2 Feature extraction  [3.Annotation]  [4.Experiments and evaluation]

3. 1.Introduction Emotional emails often contain complaints about unfair treatment due to negligence, incompetence, rigid protocols, unfriendly systems, and unresponsive personnel. Reduce the churn i.e., retaining customers who otherwise would have taken their business elsewhere.

4. 1.Introduction A negative emotional component articulated by phrases like: you suck, disgusted. Enumerating factual sentences such as: you overcharged, take my business elsewhere.

5. Outline  [1.Introduction]  [2.Emotion detection in emails] 2.1 Classifier 2.2 Feature extraction  [3.Annotation]  [4.Experiments and evaluation]

6. 2.Emotion detection in emails We use supervised machine learning techniques to detect emotional emails.

7. 2.Emotion detection in emails 2.1 Classifier We used Boostexter as text classification. The output of the final classifier f is: i.e., the sum of confidence of all classifiers The final classifier f can be mapped onto a confidence value between 0 and 1 by a logistic function;

8. 2.Emotion detection in emails 2.1 Classifier Training model:

9. 2.Emotion detection in emails 2.2 Feature extraction Salient features For our data we have identified the eight features listed below.

10. 2.Emotion detection in emails 2.2 Feature extraction

12. In the labeling guide for critical emails labelers were instructed to look for salient features in the email and keep a list of encountered phrases. We further enriched these lists by: a)using general knowledge of English, we added variations to existing phrases. b) searching a large body of email text (different from testing) for different phrases in which key words from known phrases participated. For example from the known phrase lied to we used the word lied and found a phrase blatantly lied.

13. Outline  [1.Introduction]  [2.Emotion detection in emails] 2.1 Classifier 2.2 Feature extraction  [3.Annotation]  [4.Experiments and evaluation]

14. 3.Annotation 1) Two different labelers. 2) Kappa : 0.814.

15. 3.Annotation

16. Outline  [1.Introduction]  [2.Emotion detection in emails] 2.1 Classifier 2.2 Feature extraction  [3.Annotation]  [4.Experiments and evaluation]

17. 4.Experiments and evaluation We used cross validation (leave-one-out) technique on the test set.

18. 4.Experiments and evaluation