In the previous three chapters, we see how the conformal prediction (CP) framework is adapted to handle real-world problems for face recognition, medical applications, and also network traffic/demand prediction. The theoretically proven validity property of conformal predictor makes it an extremely attractive prediction tool for many real-world prediction tasks for both the online and offline settings. Moreover, the ability of the conformal predictor to return prediction region (or interval) based on user-defined confidence level, instead of just returning a single point prediction, is an attractive capability that other conventional machine learning techniques and statistical models cannot provide. These are the two main reasons more practitioners are seriously considering using conformal predictors for their real-world applications. In this final chapter, we provide a review on other real-world applications that directly utilized the CP framework for classification/regression tasks or its adaptations to tasks not discussed in the previous three chapters.
|Original language||English (US)|
|Title of host publication||Conformal Prediction for Reliable Machine Learning|
|Subtitle of host publication||Theory, Adaptations and Applications|
|Number of pages||11|
|State||Published - Apr 2014|
All Science Journal Classification (ASJC) codes
- Computer Science(all)