'Get reasons why the model made a specific prediction
So I'm a novice working on a tensorflow algorithm that predicts whether a credit card transaction is a fraud or not by giving either 0 or 1 as an output. So far it has been working well, but I think one of the things it is missing is that it doesn't give a reason why it made a specific prediction. Ideally it should be able to tell people which specific parameter(s) caused it to classify a transaction as a fraud, so that it could be investigated further. Is there a good way to achieve this?
Sources
This article follows the attribution requirements of Stack Overflow and is licensed under CC BY-SA 3.0.
Source: Stack Overflow
| Solution | Source |
|---|