Modelarea Sintactica a Sistemelor Biologice

Ciprian Chichirita

grupa 407 (I.A.)

Predicting individual well-being through the language of social media

Contents

1. Quality of Life studies
2. Quality of Life
3. Used methods
4. Investigation means
5. The problem
6. The proposal
7. Innovation
8. Bibliography

1. Quality of Life studies

✔ The good life, BBC Survey, 1001 People
      ✔ greatest happines - 81%
      ✔ greatest wealth - 19%

✔ Other studies, an average of 69% of people globally rate well-being as more important than any other life outcome

2. Quality of Life

✔ Positive life outcomes
      ✔ health and longevity
      ✔ needs pyramid (Maslow's hierarchy of needs)

2. Quality of Life

✔ Well-being is more than simply positive emotion or mood
      ✔ meaning in life
      ✔ engagement in activities
      ✔ the state of one’s relationships
      ✔ positive emotion

✔ Well-being offers a preventative approach to public and personal health, with important economic consequences

3. Used methods

✔ Satisfaction with Life (SWL)

✔ PERMA
     ✔ positive emotions
     ✔ engagement
     ✔ relationships
     ✔ meaning
     ✔ accomplishment

4. Investigation means

✔ door to door survey
✔ online survey
✔ media

5. The problem

✔ people that are sick will not go to a doctor or will not be willing to participate in any kind of surveys

✔ medical information collected by doctors was being collected in a huge time span, like 4 to 5 years and the problem with this kind of survey is that the form that the doctor was initially filling was prone to changes, thus, at the end of the survey, we will have inconsistent data and in need to be filled with forecast values

✔ anonymity

6. The proposal

✔ predicting well-being based on natural language use
✔ a user typically writes 123 messages (in a month, on twitter and facebook)
✔ :(= negative polarity
✔ ngrams (unigrams and bigrams)
✔ topics running latent dirichlet al-location (LDA) over a set of 18 million Facebook status updates from the MyPersonality appli-cation
✔ lexica categories of words from Linguistic Inquiry and Word Count (LIWC)
✔ Message-level Models > User-Level Models > Cascaded Message-to-User Level Well-Being Prediction

6. The proposal

✔ To acquire PERMA and SWL indicators from facebook and twitter data, Amazon’s Mechanical Turk (MTurk) was used. The algorithm found out that if a post had one of these words: tonight, tomorrow, excited, super pumped, stoked, psyched, thankful, wonderful, amazing, grateful, blessed, loving, supportive, friends, lucky, it meant that the person posting was in a pretty good shape mentally and physically.

7. Innovation

✔ introduction of the task of predicting individual well-being
✔ finding of a two-level, message-to-user model to perform better than models based on either independently
✔ analysis of the linguistic features that is associated to individual satisfaction with life
✔ a well-being language model available for researchers

Bibliography

http://psb.stanford.edu/psb-online/proceedings/psb16/schwartz.pdf

https://en.wikipedia.org/wiki/Main_Page

Resources

https://upload.wikimedia.org/wikipedia/commons/3/33/MaslowsHierarchyOfNeeds.svg

https://www.google.ro/url?sa=i&rct=j&q=&esrc=s&source=images&cd=&cad=rja&uact=8&ved=0ahUKEwjbzZu14uPMAhUDlxQKHauJBz8QjRwIBw&url=http%3A%2F%2Fwww.performancemagazine.org%2Fcreate-your-better-life-index%2Flife-quality%2F&bvm=bv.122129774,d.bGg&psig=AFQjCNHBViq2J41bg17qD17rNNNeTXge_Q&ust=1463665864938768

Thank you!