Bot_ToDoList.txt

List of features to implement for Mads Poulsen


#Tweets/Retweets
- Tweeting with location (geo-tags)
- Retweet 2 good machine learning tweets a day
        1. Best probability of being good
        2. Most retweets within last 24 hours
        *  Make a rule where the above two are the same
DONE - Retweet people from San Francisco

#Data
- Collecting good tweets (tweets with more than 10 retweets)
- Collecting bad tweets (tweets with less than 10 retweets)
- Extract features from tweets
        * Followers of the user who tweeted
        * Age of the user account
        * Nr. of links in the tweet
        * Nr. of words in the tweet
        * Nr. of hashtags in the tweet


#Machine Learning
- Classification on friends (good or bad?)
- Classify tweets collected (FOLLOW from Data 'Extract features')
        * Naive Bayes Classifier
        * Maximum Entropy Classifier


#Follow
DONE - Implement ''simple'' (reciprocal) follow routine
- Implement follow routine on good friends
- Follow 50% spam accounts, 50% human accounts until 1000 followers
- Follow people (FOLLOW from Relocate 'Move to San Francisco')
        * Gain followers in San Francisco

#Relocate
DONE - Move to San Francisco