Data Analysis and Confidence based on SVM Density Estimation

Data-driven models are frequently used in industry to predict various characteristics of processes. In order to build robust model, the quality of the data needs to be analysed. These models are also required to associate a level of confidence with their predictions. In a high-dimensional setting it is important to incorporate data density information when analyzing the quality of the data and the determining the confidence in a prediction. The SVM density estimation together with results from the typicalness framework form a powerful tool that is effective for industrial applications.