Preselection of documents for personalized recommendations of job postings based on word embeddings

In the search for matching jobs, more and more people rely on online services such as job search engines. Job search engines provide the possibilities of searching for certain keywords and maintaining domain related filters like the location or the seniority of a job posting. A job recommendation system can support the users on such platforms by finding relevant jobs that match their profile. When it comes to job postings, the platform often has no information about whether a user actually applied for a certain job or whether the application was successful. In this paper, we propose a method to use the implicit information that users provide on the platform to recommend matching job postings in real time. We provide a solution by applying the doc2vec method on the job descriptions to cluster them. This allows us to preselect certain job postings and reduce the target space to implement a personalized classifier for recommendation. Both the quality of recommendations and the runtime of the according algorithms are improved. Our evaluation with domain experts shows, that at least 55% of these recommendations are relevant to the respective user.

[1]  Pasquale Lops,et al.  Learning Word Embeddings from Wikipedia for Content-Based Recommender Systems , 2016, ECIR.

[2]  D. Sculley,et al.  Web-scale k-means clustering , 2010, WWW '10.

[3]  Christoph Rensing,et al.  Multi-label Text Classification Using Semantic Features and Dimensionality Reduction with Autoencoders , 2017, LDK.

[4]  Quoc V. Le,et al.  Distributed Representations of Sentences and Documents , 2014, ICML.

[5]  Christoph Rensing,et al.  Combining Active and Ensemble Learning for Efficient Classification of Web Documents , 2014, Polibits.

[6]  Xueqi Cheng,et al.  An ensemble method for job recommender systems , 2016, RecSys Challenge '16.

[7]  Christoph Rensing,et al.  Text classification based filters for a domain-specific search engine , 2016, Comput. Ind..

[8]  Raymond J. Mooney,et al.  Content-boosted collaborative filtering for improved recommendations , 2002, AAAI/IAAI.

[9]  Michael P. O'Mahony,et al.  An Analysis Framework for Content-based Job Recommendation , 2014, ICCBR 2014.

[10]  Martha Larson,et al.  Exploring Deep Space: Learning Personalized Ranking in a Semantic Space , 2016, DLRS@RecSys.

[11]  Núria Bel,et al.  Ranking Job Offers for Candidates: learning hidden knowledge from Big Data , 2014, LREC.

[12]  Y. Huang,et al.  Exploiting Embedding in Content-Based Recommender systems , 2016 .

[13]  Ruchi Agarwal,et al.  Review of job recommender system using big data analytics , 2016, 2016 3rd International Conference on Computing for Sustainable Global Development (INDIACom).