evaluation of survey

currantcurrant Member Posts: 14 Contributor II
edited November 2018 in Help
Dear All,

in our company we conducted a survey among the employees (250 employees) . The participants could answer some questions to different topics (e.g. topic A: "communication", topic B: "Own actvities", topic C: "direct superior" (= direct boss), ... )
The possible answers (nominal) were "not relevant at all" (= 1 point), ..., "applies completely" (= 5 points).

Now, I want to find groups of employees within our company (e.g. -> satisfied employees, ...)

Up-to-now, I tried k-means: I calculated the average points for each employee within each topic and ran the k-means. here I found four groups (the best number of groups).

Are there any other approaches (PCA, ...) to find groups within the results of the survey?

Thanx in advance



  • MariusHelfMariusHelf RapidMiner Certified Expert, Member Posts: 1,869 Unicorn
    Hi currant,

    if you could calcuate the average for each employee, it means that you already transformed the nominal user input to numerical values (1-5 for each answer). On these values you could try a clustering algorithm like k-means *without* prior averaging the values. That way you use more information you gathered by your survey than with the averaging approach.

Sign In or Register to comment.