0

I am analysing some employee details to find the efficiency of the employees. Ideally I want some rankings to rank them based on these features.

My features include;

  • current salary
  • projects completed
  • customer ratings
  • number of bugs reported
  • customer complaints
  • employee satisfaction
  • profit
  • age
  • etc.

My target is the ranking of the employee in the job hierarchy.

I did some correlation analysis and found some features that have positive correlation (e.g., salary, projects completed, profit, customer ratings, employee satisfaction etc.).

I also found some features that are negatively correlated with ranking (e.g., customer complaints, number of bugs reported etc.)

Now, I want to use these features to make a function that reflects the employees actual rankings, so that we can use the same function to rank employees in the future.

NOTE: I am not only limited to correlations. I am happy to receive answers using other methods as well.

Is there a any good method to solve my problem.

I am happy to provide more details if needed.

EmJ
  • 592
  • 3
  • 15
  • Please indicate measurement- types numeric for different "features" you want to incude. in your measure of efficiency. how do you compute efficiecy? –  Apr 18 '19 at 04:27
  • @SubhashC.Davar I want some ranking (i.e. the most efficient employee to be 1 and so on). Do you have any suggestions? :) – EmJ Apr 18 '19 at 04:52
  • kindly respond to clarificatios if you want an approrate answer from the users of this accredoted site. –  Apr 18 '19 at 07:40
  • I guess principal component factor analysis will be used somewhere. Based on those variables, run a PCA and use the first loading to rank the students. – user245165 Apr 18 '19 at 03:37
  • 1
    Do you have an external measure of the efficiency of a worker? If not, do you know in advance which "pole" of _every_ of your features is positive ("efficient")? – ttnphns Apr 18 '19 at 20:43
  • @ttnphns No, I do not have any external measures of workers. In most of my features highest value assume to be more efficient. However, there are some features such as `customer complaints, number of bugs reported` which are better when the employee has a lower score. Do you have any suggestions? Looking forward to hearing from you. Thank you very much. – EmJ Apr 18 '19 at 21:55
  • As already suggested to you, that seems to an undertaking of factor analysis or at least PCA. That will allow you to construct few (maybe one or two or three) latent constructs (dimensions) of "efficiency", the kinds of efficiency. If your features are not all interval scale but some are categorical, then this is a problem, but you could try CatPCA or other approaches ([read](https://stats.stackexchange.com/q/215404/3277)). You should first read literature on these kind of analysis for it is not quite easy. – ttnphns Apr 19 '19 at 07:02
  • 1
    It sounds like you might already have known target ranks for each employee. Is this the case? And, are you trying to learn a ranking function that matches these targets? If that's the case, could you edit the question to make this clear? – user20160 Sep 08 '19 at 04:36
  • @user20160 Thank you for the comment. Yes, I do have target ranks for this year. However, our intention is to make a ranking function that we can use in future (more specifically yearly) to rank the employees. Please kindly let me know if you have any suggestions :) – EmJ Sep 08 '19 at 05:28

0 Answers0