Explanations are essential components in the promising fields of artificial intelligence (AI) and machine learning.  Deep learning approaches are rising due to their supremacy in  terms of accuracy when trained with huge amounts of data. Because of their black-box nature, the predictions are also hard to comprehend, retrace, and trust.  Good explanation techniques can help to understand why a system produces a certain prediction and therefore increase trust in the model. Understanding the model is crucial for domains like healthcare, where decisions ultimately affect  human life.  Studies have shown that counterfactual explanations  in  particular  tend  to  be  more  informative  and  psychologically  effective than other methods.

This work focuses on a novel instance-based technique called “Native Guide”, that  generates  counterfactual  explanations  for  time  series  data  classification. It  uses  nearest  neighbour samples  from  the  real  data  distribution  with  class change  as  a  foundation.   This  thesis  applies  the  method  on  the  explanation of  electrocardiogram  (ECG)  classification,  a   very  complex  and  vital  medical field where every single ecg carries unique features.  Native
Guide for ECGs  is explained, examined and expanded by providing necessary background knowl edge, amplifying aspects like plausibility, comparing different suitable models to  each  other  and  indicating  benefits  and  downsides.   Finally,  counterfactual explanations for ecg data classification generated by Native Guide are evaluated by cardiologists by means of two expert interviews.

Synchronization of the periodic ecg data was shown to be the most important contribution  to  the  method  that  enabled  the  generation  of  plausible  coun- terfactuals.  The experts, who had never seen or used counterfactuals in their work, were interested in this approach and could envision its application within the  field  when  it  comes  to  training  junior  doctors.   In  general,  AI  classification  along  with  sophisticated  proximate  counterfactuals  indicate  success  and reliability when it comes to the identification of heart diseases.