1st Data Set: Predicting Solar Plant Energy Output
-Problem: Because solar power plants are at the mercy of weather forecasts, it is difficult for a solar energy plant to accurately calculate how much energy it can generate.
-Solution: Using predictive analytics, we can predict how much energy we can produce more efficiently.
-Results: By being able to predict how much energy a solar plant can generate in the future, the company can work more efficiently.
2nd Data Set - Predictive Maintenance for Wind Turbines
-Problem: Wind turbines have many factors that affect the health of the entire system, such as air temperature, torque, or wind speed. Because these factors are out of our control, there's a chance that wind turbines can break. When they break, repairs can be costly.
-Solution: To avoid downtime costs, a wind power company can use predictive analytics to predict the probability of machine failure and proactively maintain their wind turbines.
-Results: By using predictive analytics, a company can have a concrete maintenance plan to reduce downtime and eliminate costly repairs.
Minitab
Predictive Analytics Webinar Series
Leverage the power of predictive analytics with Minitab.
There are many valuable algorithms that provide deeper insights into your data. Watch our on-demand webinar series to explore our best-in-class Machine Learning algorithms.
Minitab’s Predictive Analytics Webinar Series introduces you to our popular algorithms:
- CART®
- Random Forests®
- TreeNet®
- MARS®
Using our most advanced predictive analytics solutions, we’ll dive into two real-life data sets and highlight how Minitab can be used to predict solar energy output and prevent maintenance in industrial machinery.
- Example 1: How to Predict Energy Output of a Power Plant
- Example 2: Predictive Maintenance for Wind Turbines
Webinar 1
CART & Random Forests
In this webinar, we will cover Classification and Regression Trees (CART®), a powerful approach that solves predictive analytics problems in a data-driven way. We’ll explain the rationale behind the algorithm, its unique advantages, along with two real-life examples of solar and wind power applications. Lastly, we will introduce Random Forests algorithm as an ensemble of CART trees to increase the accuracy of the model.
Webinar 2
TreeNet
In this webinar, we continue our investigation of the modern tree-based predictive analytics methods by introducing the stochastic gradient boosting algorithm – TreeNet®. We highlight some of its history, its core design principles, areas of applicability, as well as unique advantages that achieve superior predictive accuracy.
.png)
Webinar 3
MARS
In our final webinar of the series, we will introduce the newest PA tool available at Minitab – MARS®. Designed to bridge the gap between classical multiple linear regression and modern data-driven algorithms such as TreeNet® and Random Forests®, MARS automatically discovers possible non-linearities in your data and packages them into a neat and easy-to-present series of equations. We’ll demonstrate the power of MARS for regression problems, as well as binary classification problems arising in solar and wind power applications. If you want to outperform traditional regression techniques while preserving the simplicity of your models, MARS is the tool of choice.
Data Sets
