Google Prediction API

Google’s machine learning algorithms w/ RESTful interface

#5 Product of the WeekApril 11, 2015
There are no images or videos added to the gallery.
Add to gallery



You need to become a Contributor to join the discussion - Find out how.
Thomas R Stone
@thomasrorystone · Co-founder, PredictionIO
After seeing AMZN Machine Learning yesterday surprised this wasn't posted. Interested to hear people's experiences with this and also similar offerings from MSFT ( and AMZN ( We spent a lot of time building open source (Apache 2) PredictionIO ( to make it useful for real-world production usage not just toy ML problems.
Mike Coutermarsh@mscccc · Code @ Product Hunt
@tstonez Hey Thomas. What are top real-world use cases you've been seeing? ~200ms response time seems almost too good to be true, how do the different services stack up? Are the API and speed key differences?
Thomas R Stone
@thomasrorystone · Co-founder, PredictionIO
@mscccc can't speak for GOOG but for us at it seems Recommender Systems are still one of the most popular use cases, perhaps because Mahout was the 1st library we supported, but now PredictionIO integrates MLlib, Deeplearning4j, CoreNLP we see all kinds of use cases such as Demand Prediction, Churn Analysis, Fraud Detection, Lead Scoring, Documentation Classification, Sentiment Analysis …and more. Query response time is important but also the ability to add custom business logic, data security, flexibility to modify different parts of the ML pipeline, or "engines" in our parlance, and last but not least a strong community. Be interesting to see how the various MLaaS offerings from MSFT/GOOG/AMZN develop but on many of these points I believe open source wins.
Tom MasieroHiring@blendahtom · GM of LaunchBit
@tstonez @mscccc I'm with you Mike.. would love to understand some real world use cases.
Federico Pascual@federicopascual · COO @MonkeyLearn.
Google Prediction API is a great platform. If you are interested in machine learning, you should also check out MonkeyLearn, a text mining platform, where developers can get data from text using machine learning. One of the things that set us apart from Google Prediction API, is that besides using our API for common text mining tasks like topic classification or sentiment analysis, you can create your own custom text classifier, trained with your own data, that fit your needs. We also have a nice UI where you can play and build custom algorithms without any knowledge in NLP or machine learning. You can learn more here: