OWI-lab helps companies through the process of applying and integrating data mining methods for condition monitoring in order to complement or improve efficiency and accuracy of existing monitoring systems. Data mining in general helps to discover new knowledge about a system or process. The data can consist of continuous data, such as time series of sensor measurements with or without temporal reference as well as discrete data such as event messages.
Examples of data mining methods are clustering (unsupervised), classification (supervised), regression (supervised) using – amongst other - a decision tree, neural network, or support vector machine models. Whereas clustering groups objects in such a way that objects in the same group/cluster are more similar to each other than to those in other clusters, classification identifies to which category a new observation belongs. Regression on the other hand estimates or predicts a continuous response.
Data mining can for example be used for fault detection, fault localization, life time estimation of components, anomaly detection or decision support. OWI-Lab can help companies with all key steps within the data mining process, i.e. data cleaning, preprocessing, feature extraction, choosing the right model, training the model, and analyzing and presenting the results. This way, we will help you reach accurate and impressive results by utilizing the information you already have, as well as formulate strategies for even greater improvement.