Skip to main content

Hitachi
Contact UsContact Us

Analytics Platform for R&D

Hitachi provides cloud-based material data analytics platform.
Simple UI to operate Hitachi’s original AI powered tools easily on a web browser.

[image]Overview of Analytics Platform for R&D

FUNCTIONS

LEARN

Build prediction models with proven machine learning algorithm which you can choose from 8 options, or an algorithm developed on your own.

PREDICT（FORWARD ANALYSIS）

Select an algorithm from multiple choices commonly used for machine learning and Materials Informatics. Predict experimental results by entering experimental conditions that match the prediction model you created.

[image]Overview of PREDICTION
Screen image for illustrative purpose only.

・Multiple algorithm choices commonly used in machine learning including:
- Random Forest
- Gradient Boosting
- Lasso
- etc.
・Include popular algorithms in Materials Informatics such as:
- Gaussian Process Regression
- PLS
- AD-score
- etc.

OPTIMIZE (INVERSE ANALYSIS)

Output experimental condition which the prediction model estimated as optimized condition to derive the result you target.

  • Set multiple objective valuable.
  • Weight importance of objective valuables.
  • Constrain conditions for explanatory valuables.
  • Select an optimization algorithm from multiple choices. (Mathematical, Bayesian optimization etc.)

VISUALIZE

Visualize analysis results with highly flexible visualization tools.

[image]Overview of VISUALIZATION
Screen image for illustrative purpose only.

  • Scatter plots with changeable plot color and plot size.

DEVELOP YOUR OWN MODELS

Provide your own algorithm with a Python environment in the cloud.

[image]Overview of CODING YOURSELF
Screen image for illustrative purpose only.

  • Easy to deploy and share Python codes for specific analysis created by data scientists in your organization.