Considering the answers you found for the topic of games/news in the last assignment, you will group by
topic & age, topic & country, topic & gender as well as topic & profession and find out the pattern and breakdown for each of those groups. Consider those categories for each of the groups:
a.share w/media
b.share w/o media
c.fake news (t & f)
d.time spent w/ media
e.time spent w/o media
You will calculate the p-value for each one and formulate a null hypothesis. Can you find any pattern in each one of those aggregated groups? The only significant ones will be the samples of at least 29 members on each group
You need to create as many regression models as possible and display well the results. There are a few different targets:
4a. Share w/ media
4b. Share w/o media
4c Fake news
4d. Share Fake news
4e. Share True news
4f. Time Spent with media (linear)
4g. Time Spent w/o media (linear)
Both 4f & 4g has combinations with fake or not fake news that can be all explored for the final project
For the ones who did not calculate statistical significance for the games cultural values, now is the time to do so. Make sure to write your hypothesis
- Try a classification tree model for at least two of those targets:
4a. Share w/ media
4b. Share w/o media
4.c Fake news
4d. Share Fake news
4e. Share not fake news
Train the decision tree model only the numeric attributes, binned age, binned profession. See my recording lecture (12/10)
- Compare the decision tree results & performances with the equivalent regression model.
This is just a conceptual exercise, we know that the number of records is not enough to train a classification tree model (one needs at least 3 thousand records).
Comments from Customer
The reason why the username may appear multiple times is because when the user plays a game whether it's a sport, education or entertainment, etc, their username is linked to that game. The user creates an account so every game that they play when they're logged in will be associated to their username. This is the only file that was provided by the professor and it's the cleanest form. Please let me know if you have any questions