I am new to data science and I am hoping I can start applying it to my job. I have watched some videos on places like Udemy for Machine Learning and Python etc.

Anyway, I have a task but I am not sure what I need to look into to solve it. I have a PCB/Circuit board which has 80 components on it. I also have a piece of test equipment that measures each component and logs the results with a reading of what was measured and then a simple 1 or 0 for the result I.E. Pass/Fail

Lately we have notice a component has started to fail a lot on the tester. One of the tasks I have been give is has any of the other components changed in their readings and could this have an effect?

So my question is, what would I use if I wanted to see if another component potentially has influence?

An example may help if my explanation is poor.

Lets says I have these components and the general readings I expect

Component    Reading     Outcome
C1            1.5 nf     Pass
C2            1.0 uf     Pass
C25           10 nf      Pass
R22           100 kohms  Pass 
R53           150 kohms  Pass

Now lets say C25 is failing because the reading has gone up.

Component    Reading    Outcome
C1            1.5 nf    Pass
C2            1.0 uf    Pass
C25           12 nf     Fail
R22           80 kohms  Pass
R53           150 kohms Pass

If you notice, R22 has dropped. Now say that this is the pattern. If R22 reads low then C25 can read high and potentially fail. What can I use to find this link/pattern/relationship?


You could try running a stepwise logistic regression with C25 pass/fail as the response and the readings of the other components as predictors.


I would look at this answer here.

Statistically speaking, there may be a correlation if you see variables moving together (either towards or opposite) which can help you understand the relationship. You would want to use the Readings column here.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.