Practical Application to Interpretable Medical Risk Analysis

Attributes of weighted fuzzy rule interpolation (FRI) techniques have seen successful applications over a wide range of applications, as demonstrated in the preceding chapters, covering tasks such as pattern recognition, classification, and prediction. This chapter presents a systematic application in the carefully selected medical domain, addressing the challenging problem of mammographic mass risk assessment. First, it introduces the background regarding this domain problem, motivating the real-world case study. Then, it describes the mammographic image data considered, followed by an exhibition of the fuzzy rule-based interpolative reasoning process in deriving interpretable outcomes on mass risk assessment. The chapter also presents a statistical analysis of the application results, offering a concrete example for gauging the success (or otherwise) of applying the key approximate knowledge-based reasoning techniques introduced in this book to solve real-life problems.