As a first step into error prediction, we can take the following approach:
- define a set of input conditions
- run the application with the given conditions n times
- define the probability of success / failure based on the frequency of each outcome
This could be useful because some conditions will fail every time because they are invalid, while other conditions may only fail sometimes depending on the system or random chance (or something like that). So it would be good to be able to predict the probabilty of failure based on historical data.
As a first step into error prediction, we can take the following approach:
This could be useful because some conditions will fail every time because they are invalid, while other conditions may only fail sometimes depending on the system or random chance (or something like that). So it would be good to be able to predict the probabilty of failure based on historical data.