Predicting Success

With this line of work we aim to forecast which products will be successful. Then we would like to use this information to answer questions about where and how to optimally allocate resources to take advantage of the future success.

Basic Questions

Is there information in the precursory dynamics that can predict future success?
Is there a way to aggregate many sources of information to build a picture of what is becoming popular?

Acceleration

Develop techniques to measure the acceleration of interest, and quantify the extent to which this information can be used to predict future success.

Spidyn

In addition to using the power law for foreshocks, there is a system called spidyn which predicts acceleration upside as well as downside that is quite successful in trading. This could be used to study the YouTube time series to predict future success because it is a very robust system which uses the same qualitative ideas. Furthermore the spidyn system will work more robustly than power law fitting for the prediction of acceleration and future successes.

YouTube Data

We are currently analyzing YouTube data.

New Data

New data and ideas of clean experiments are needed for this branch of research far more than any other branch.

JavaScript has been disabled in your browser