Predictive Models for Behavioral Outcomes through Crowdsourcing

To develop models from large datasets and determine which subset of data to mine is now becoming automated. Choosing what kind of data to be collected and place it first, requires an immediate apprehension by human which is supplied by domain expert. A new approach to machine science which demonstrates that even non domain experts can express clearly and precisely the features and provide values for those features so that they are predictive of some behavioural outcome of interest. It is done by creating a web platform where interaction between groups of people by responding to questions which help to predict behavioural outcome, results in dynamically growing online survey. This leads to predict the behavioral consequences of user with the help of their responses to survey questions formed.