Kaggle Competition

P123 has paid to have an AI specialist look at machine learning. P123 got some data on Random Forests.

Kaggle is a competition for machine learners. Thousands compete for prizes.

Prizes range from Kudos or $10,000 in monetary prizes to $1,500,000. Link to prizes page HERE. Notice the $10,000 prize got a lot of competitors. Also notice some people value this at $1,500,000 (“Passenger Screening Algorithm Challenge Improve the accuracy of the Department of Homeland Security’s threat recognition algorithms”)

Kaggle became famous after a winning algorithm for detecting the Higg’s Boson was adopted by CERN. There have been competitions for Investment Firms where the prize was employment to further develop the winning algorithm.

If P123 still has an interest in understanding the potential for machine learning this may be the place to look.

Basically P123 would provide the training set that they gave the AI specialist. With masked factors/functions ie., f1, f2, f3, f4,…f50

The winner is the team who’s trained algorithm does the best with the holdout set (again could be masked). Like what the AI specialist looked at but with over a thousand machine learners motivated to actually find something that works.

Back in the day, the competitors would try a Random Forest (now and then).

If this fails P123 would know to close up shop or just provide fundamental data for fundamental analysts.

If there is a successful winning algorithm P123 would now have a new working algorithm and would not know what to do with the flood of new members.

Notice the competition is both famous and well respected. A winning algorithm would be advertisement for P123. The Department of Homeland Security has offered a prize of $1,500,000 from the Department of Homeland security which is testimony to the respect and effectiveness, I think.

BTW, some of these teams are genius graduate students–from many countries–working on Mac Laptops (Python usually). So, computer resources for the winning algorithm may not be a problem.

Anyway, if you do this I want to enter. Enter with my second best algorithm. Of course, I will probably abandon my best algorithm for the algorithm that wins the competition. I am pretty sure 1,000 of the best machine learners could beat my best algorithm.

[b]FWIW, I am here for what Marco does. I do not pay a cent for Marc or Yuval’s advice. To the extent Marc helps with data I am glad he is here. If Yuval helps advertise the platform over on SeekingAlpha I am glad he is here too.

[color=firebrick]Marco needs to expand the capabilities of the platform. Not limit the usefulness.[/b][/color]

I think the Kaggle competition should be looked at for answers regarding how Marco might want to expand the P123 platform…

Edit

Examples:

For Marc: March Machine Learning Mania 2016 Predict the 2016 NCAA Basketball Tournament $25,000

Two Sigma Financial Modeling Challenge Can you uncover predictive value in an uncertain world? $100,000

Two Sigma: Using News to Predict Stock MovementsUse news analytics to predict stock price performance Prize: $100,000

Zillow Prize: Zillow’s Home Value Prediction (Zestimate) Can you improve the algorithm that changed the world of real estate? Prize $1,200,000

Data Science for Good: City of Los AngelesHelp the City of Los Angeles to structure and analyze its job descriptions. Prize $15,000

Microsoft Malware Prediction Can you predict if a machine will soon be hit with malware? Prize $25,000

Google Analytics Customer Revenue Prediction Predict how much GStore customers will spend Prize $45,000

Home Credit Default Risk Can you predict how capable each applicant is of repaying a loan? $70,000

Recruit Restaurant Visitor Forecasting Predict how many future visitors a restaurant will receive $75,000

Predicting Red Hat Business Value $50,000

CHECK OUT THE LINK!!!

-Jim