Xomnia developed a solution for Randstad to determine if someone’s job loyalty can be predicted based on a resume. This is being done in 3 steps. In step 1 CV’s (with a random file format) are being converted to one uniform file format. In step 2 an algorithm extracts a broad range of features from each CV. This is being done on two dimensions: features related to the style of the document (e.g. number of spelling errors, word diversity, word count) and features related directly to the person. Think about someone’s educational or professional background and general information (such as age, gender, etc). In step 3 machine learning techniques are being applied and benchmarked on a historic dataset to test the predictive value of these features towards loyalty. We proved that there is predictive value hidden in these CV’s and most importantly, that it is possible to extract a new recruitment feature: a job loyalty indicator.
"The data of Randstad is a strategic asset. Xomnia enables us to unleash its business value"