Predicting Heart Attack through Explainable Artificial Intelligence

EasyChair Preprint no. 2093

Abstract

A novel classification technique based on combined Genetic Algorithm (GA) and Adaptive Neural Fuzzy Inference System (ANFIS) for diagnosis of heart Attack is reported. Exploiting the combined advantages of neural networks, fuzzy logic and GA, the performance of the proposed system is investigated by evaluation functions such as sensitivity, specificity, precision, accuracy and Root Mean Squared Error (RMSE).  Also, the efficiency of the algorithm is evaluated by employing 9-fold cross validation. To address the explainability of the predictions, explainable graphs are provided. The results show that the performance of the proposed algorithm is quite satisfactory. Furthermore, the importance of various symptoms in diagnosis of heart attack is investigated through defining and employing an importance evaluation function. It is shown that some symptoms have key roles in effective prediction of heart Attack.

Keyphrases: Artificial Intelligence, Explainable Artificial Intelligence, GA, Heart Disease

@Booklet{EasyChair:2093,
  author = {Mehrdad Aghamohammadi and Manvi Madan and Jung Ki Hong and Ian Watson},
  title = {Predicting Heart Attack through Explainable Artificial Intelligence},
  howpublished = {EasyChair Preprint no. 2093},

  year = {EasyChair, 2019}}