infsoft Machine Learning
infsoft Machine Learning is a visual tool that allows creating user-defined machine learning models, train them within a very short time and use them in a wide variety of applications. The powerful environment processes position and/or sensor data and uses self-optimizing algorithms that can learn from experience. By recognizing patterns and regularities in existing data, values and results can be predicted. Gained insights can be utilized as a basis for decisions and feedback mechanisms with regard to previously unknown situations.
The data from infsoft Machine Learning can also be used in third-party systems via Web Services.
The AI can be trained on the basis of different algorithms. Depending on the application, a neural network is selected for the training. In addition, filters can be stored in order to, for example, exclude irrelevant data. infsoft Machine Learning offers users a clear drag-and-drop interface for this.
Models can be used to automatically capture certain values (e.g. asset attributes). This data can then be used in the infsoft Processing & Output tools (e.g. for analytics functionalities, for actions in the Automation Engine, for the assignment of asset attributes in infsoft Tracking, or for work-sharing processes in infsoft Workflow Management). The values can also be made available to third-party systems via web services.
When data sets are stored, processed, and analyzed using intelligent algorithms, the system can predict values for the future. For example, infsoft Machine Learning can be used to make predictions about the location of an asset, the amount of carbon dioxide in the air, the number of visitors in a shopping mall, or waiting times at an airport at a given point in time. In industrial environments, reliable statements can be made about the condition of machines and systems, enabling expected malfunctions to be recorded and corrected as quickly as possible.
infsoft Machine Learning delivers part of the data as scores (validation data). These scores can be used to evaluate the data provided by the models. This makes it possible to determine whether the predictions are accurate and how high the error rate is. If necessary, adjustments can be made to the algorithm to obtain even more accurate results and select the final methodology.