Determining People’s Emotions in Facebook

Determining People’s Emotions in Facebook

Determining People’s Emotions in Face book management report in Python. Opinion Mining and Emotion Mining are part of the Sentiment Analysis area, but they have different objectives. Opinion Mining is concerned with the study of opinions expressed in texts and its basic task is polarity detection, whereas Emotion Mining is related to the study of emotions and its basic task is emotion recognition. Polarity detection is usually a binary classification task with outputs such as positive vs. negative or like vs. dislike, while emotion recognition aims to enable computers recognize and express emotions. Click here to get complete Python projects lists.

In this paper we focus on Spanish emotion classification. We first compile a corpus from Facebook using the reactions in comments and posts in order to label different emotions. Then we apply a basic machine-learning approach and two lexicon-based approaches, one using a Spanish version of the NRC Emotion Lexicon (Emolex) and another adapting WordNet-Affect to Spanish. The results demonstrate the difficulty of the task and show some interesting features in the lexicon approaches.

Conclusion

Determining People’s Emotions in Face book management report in phython.We have explored the potential of using Facebook reactions in a distant supervised setting to perform emotion classification. The evaluation on standard benchmarks shows that models trained as such, especially when enhanced with continuous vector representations, can achieve competitive results without relying on any handcrafted resource. An interesting aspect of our approach is the view to domain adap tation via the selection of Facebook pages to be used as training data. We believe that this approach has a lot of potential, and we see the following directions for improvement. Determining People’s Emotions in Facebook Report.

Feature-wise, we want to train emotion-aware embeddings, in the vein of work by Tang et al, and Iacobacci et al. Retrofitting FB-embeddings trained on a larger corpus might also be successful, but would rely on an external lexicon. The largest room for yielding not only better results but also interesting insights on extensions of this approach lies in the choice of training instances, both in terms of Facebook pages to get posts from, as well as in which posts to select from the given pages. For the latter, one could for example only select posts that have a certain length, ignore posts that are only quotes or captions to images, or expand posts by including content from linked html pages, which might provide larger and better contexts .

Please use the link below for international payments.

Pay and Download Now