Before making changes to a device or process, you want to know the impact of that change. A predictive model, that simulates different scenarios, can give you that information. By capturing information about the impact of different machine settings and configurations, a decision support system can provide recommendations to the user, enhancing the user experience by making the product smart and intuitive.
Decision making with support of machine learning helps the user to do the right selection for preferable device performance. With the Ekkono SDK you can create a predictive model representing the behavior of the device, and use the model as a simulation tool for evaluating the effects of alternative settings and configurations, giving the user recommendations on how to best operate the device for preferred performance
Predictive models can be trained with batch or incremental learning. Batch learning is carried out offline, on already collected data, that needs to include a variety of operating conditions for the device. Proper data for batch learning can in many cases be hard to obtain, due to time and cost constraints. Incremental learning eliminates the need for historical data, by performing the actual learning in real-time on the device. Training on the device makes it possible for the model to learn the unique conditions for each device, making the model’s predictions specifically tailored to the device’s unique conditions.
Operators may not understand the consequences of their actions and may hence take suboptimal decisions.
Train a model that can explain the effect of user behavior and simulate alternative scenarios using the model.
Continuously rerun simulations and present alternative scenarios to the user.