A sentiment index measures the average emotional level in a corpus. We introduce four such indexes and use them to gauge average “positiveness” of a population during some period based on posts in a social network. This article for the first time presents a text-, rather than word-based sentiment index. Furthermore, this study presents the first large-scale study of the sentiment index of the Russian-speaking Facebook. Our results are consistent with the prior experiments for English language.
Engler and Prantl system of classification in plant taxonomy
Sentiment Index of the Russian Speaking Facebook
1. Introduction Social Sentiment Index Results
Sentiment Index of the
Russian Speaking Facebook
International Conference on Computational Linguistics
Dialogue 2014, Bekasovo, Russia
Alexander Panchenko
Digital Society Laboratory LLC, Moscow, Russia &
Universite catholique de Louvain, Louvain-la-Neuve, Belgium
June 8, 2014
Alexander Panchenko Sentiment Index of the Russian Speaking Facebook
2. Introduction Social Sentiment Index Results
Motivation: Can we measure happiness?
Why such a wild idea as measuring happiness?
Alexander Panchenko Sentiment Index of the Russian Speaking Facebook
3. Introduction Social Sentiment Index Results
Happiness: Google
Alexander Panchenko Sentiment Index of the Russian Speaking Facebook
4. Introduction Social Sentiment Index Results
Happiness: semantic field
Alexander Panchenko Sentiment Index of the Russian Speaking Facebook
5. Introduction Social Sentiment Index Results
Happiness: semantic field
Alexander Panchenko Sentiment Index of the Russian Speaking Facebook
6. Introduction Social Sentiment Index Results
Happiness: self-reported vs estimated
self-reported
are you happy now?
estimated from socio-economical factors
gross national product
quality of life, etc.
estimated from linguistic data
positive statements vs negative statements
several researchers considered this problem during last 10 years
blogs: Mihalchea and Lui (2005), Mishne and Rijke (2006),
Balog and Rijke (2006)
social networks: Dodds et al. (2011), Kamvar and Harris
(2011), Facebook Gross National Happiness index (FGNH) of
Kramer (2010)
Alexander Panchenko Sentiment Index of the Russian Speaking Facebook
7. Introduction Social Sentiment Index Results
Contributions
Two contributions of this paper to the exploration of
sentiment/happiness indexes are as follows:
We propose four indexes that gauge emotional level of
population. Unlike most previous works, we not only deal with
word-based indexes, but also device a text-based index.
To the best of our knowledge, this is the first study of
sentiment index of the Russian-speaking social network.
Alexander Panchenko Sentiment Index of the Russian Speaking Facebook
8. Introduction Social Sentiment Index Results
Facebook Corpus
Statistics of the Facebook corpus
Alexander Panchenko Sentiment Index of the Russian Speaking Facebook
9. Introduction Social Sentiment Index Results
Facebook Corpus
Number of texts per day in the Facebook corpus
Alexander Panchenko Sentiment Index of the Russian Speaking Facebook
10. Introduction Social Sentiment Index Results
Sentiment Detection Method
Most frequent positive and negative adjectives in the
Facebook corpus
1500 emotional adjectives (600 negative, 900 positive)
Alexander Panchenko Sentiment Index of the Russian Speaking Facebook
11. Introduction Social Sentiment Index Results
Sentiment Detection Method
Dictionary Sentiment Text Classification Approach
emotion delta ∆ ∈ [−1; 1] – direction and amplitude of text
sentiment
D+ – positive terms
D− – negative terms
w – word
t – text
n – length of text
α ∈ [0; 1] – a threshold
Alexander Panchenko Sentiment Index of the Russian Speaking Facebook
12. Introduction Social Sentiment Index Results
Sentiment Detection Method
Comparison of Sentiment Classification Approaches
ROMIP 2012 – Sentiment Analysis Collection
Alexander Panchenko Sentiment Index of the Russian Speaking Facebook
13. Introduction Social Sentiment Index Results
Social Sentiment Indexes
Sentiment Indexes
The goal of a sentiment index is to measure positiveness of social
network users during some period.
Alexander Panchenko Sentiment Index of the Russian Speaking Facebook
14. Introduction Social Sentiment Index Results
Social Sentiment Indexes
Sentiment Indexes
Alexander Panchenko Sentiment Index of the Russian Speaking Facebook
15. Introduction Social Sentiment Index Results
Results of the social sentiment index calculation during the
entire period
Alexander Panchenko Sentiment Index of the Russian Speaking Facebook
16. Introduction Social Sentiment Index Results
Evolution of the word sentiment index sw during the entire
period
Alexander Panchenko Sentiment Index of the Russian Speaking Facebook
17. Introduction Social Sentiment Index Results
Word (sw ) and text (st) sentiment indexes
Alexander Panchenko Sentiment Index of the Russian Speaking Facebook
18. Introduction Social Sentiment Index Results
Positive and Negative Events
(1) 31-12-*—New Year (+);
(2) 14-02-*—St.Valentine’s day (+);
(3) 23-02-*—Man’s day (+);
(4) 08-03-*—Woman’s day (+);
(5) 09-05-*—Victory Day, World War 2 commemorative day (–);
(6) 07-07-2012—Krasnodar Krai floods in Russia 6 (–);
(7) 22-07-2012—A new unpopular law regulating non-profit
organizations in Russia 7 (–);
(8) 16-09-2012—A mass protest against government in Russia 8 (–);
(9) 25-10-2012—Hurricane Sandy in US (–).
Alexander Panchenko Sentiment Index of the Russian Speaking Facebook
19. Introduction Social Sentiment Index Results
Word (ew ) and text (et) emotional indexes
Alexander Panchenko Sentiment Index of the Russian Speaking Facebook
20. Introduction Social Sentiment Index Results
Word and text emotional indexes
Word (ew ) and text (et) emotional indexes, compared to the
number of words.
Alexander Panchenko Sentiment Index of the Russian Speaking Facebook
21. Introduction Social Sentiment Index Results
Thank you! Questions?
Alexander Panchenko Sentiment Index of the Russian Speaking Facebook