H2O

What do you like best about H2O?

Excellent support for commercial product Driverless AI. Rapid iteration. Performance is generally better than one can be achieved in code.

What do you dislike about H2O?

Actually nothing. The combination of proprietary and open source tools, Driverless AI and H2O, provide tools across a full range of use cases.

Recommendations to others considering H2O:

Take advantage of the 30 day trial.

What problems is H2O solving and how is that benefiting you?

We work in both financial services and biological research.

What do you like best about H2O?

The web front end known as flow is really easy to use. It can be use to quickly create machine learning models.

What do you dislike about H2O?

The complex machine learning model overfit the data. This is especially true when the data set is small.

What problems is H2O solving and how is that benefiting you?

Trying to forecast prices using the regression models. It's very quick to test out new models.

What do you like best about H2O?

They developed top-quality open source tools, including the H2O-3 and AutoML families. I do not have a license for their Driverless AI, but my experience with it through tutorials and other demos has been superb. I should mention that their efforts to develop frameworks for ML interpretability are spot on, and their learning center is shaping up as a valuable resource to the community in general. The interfaces with R and Python enable a smooth transition of pre-existing workflows into the H2O framework.

What do you dislike about H2O?

Somewhat cryptic debugging msgs in H2O-3. They support specific packages for manipulating data (data.table in R, datataable in Python) for the sake of speed and big data maneuverability, although many users may find this limiting. Driverless AI may not be affordable to the small fish in the pond.

Recommendations to others considering H2O:

One should leverage all the resources available to test their products before buying.

What problems is H2O solving and how is that benefiting you?

I have mostly used their AutoML to build quick ML/AI prototype solutions in different domains.

What do you like best about H2O?

Easy to use with good UI design and automated ML function. Driverless AI has strong capability on the auto feature engineering and system visualization. The auto feature engineering has supported different machine learning algorithm (Random Forest, Decision Tree, Neural Network, Deep Learning, etc.) and feature parameter tuning (accuracy, time, system computing etc.) The system also helps user to reduce time and efforts for hyparemeter tuning and compare the model with different settings. This will optimize the process and provide the most efficient model for prediction in classification or regression domain. Besides, Driverless AI also has good UI design and visualization. The UI also supports end user to quickly import data, visualize data in different categories, as well as check on the system running and performance during the Auto ML process. The end user could also observe experiment summary and accuracy matrix, as well as model comparison in term of accuracy.

In addition, Driverless AI also supports the AI Interpretation to explain on the model and performance. This function is very helpful to end user for understanding the Machine Learning blackbox, as well as management team for decision making based on extensive information.

What do you dislike about H2O?

It is great if Driverless AI could support deployment for edge computing, which is common in IoT world. The edge computing will require efficient computing and good accuracy with AutoML algorithm. This will help much the customer for deployment.

Recommendations to others considering H2O:

Yes, H2O.ai DriverlessAI is a good AutoML platform for company, which does not have Data Scientist.

What problems is H2O solving and how is that benefiting you?

Data analytics for predictive maintenance. It could make easy-to-use for system operator.

What do you like best about H2O?

AutoML is a great product. They have other ones, but AutoML is the most impressive.

What do you dislike about H2O?

Nothing, maybe the integration with Spark could be improved but only in little details.

Recommendations to others considering H2O:

It's an amazing set of tools, easing the use of machine learning and giving a lot of advances to write & discover models.

What problems is H2O solving and how is that benefiting you?

Forecasting models and finding for best algorithms for my datasets