A majority of the literature in the field focuses on processing job applications. As mentioned above, we want to research the possibilities of finding expertise, and potential candidates that do not necessarily apply to the respective firm. Our research will therefore involve collecting a suitable dataset from social media sites such as linkedin and twitter. From this data we will separate out a portion that we will use for verification. The rest will be used in the development/training of the system.
Process
The derivation that forms the research questions has its origin in the motivation to be able to create value for a company, in a field that can be solved with interesting technology such as artificial intelligence and web. The research plan will be composed of two strategies. We will design an experiment during the fall semester that we later will use to evaluate the end product. During the spring semester we will follow a standard design and creation strategy while creating an AI based recruitment module. The motivations for proceeding with this strategy are several. This strategy is highly relevant in both the creation of missing knowledge and solving problems, by producing and applying knowledge to create effective artifacts. The design and creation strategy also focuses on the risky and uncertainty areas, and is a mainstay when it comes to research regarding improvement, invention and exaptation.
The experiment that will be designed during the fall semester will be used to evaluate the program’s ability to rank candidates. To achieve this, we propose obtaining a dataset from linkedIn and getting this dataset ranked by a number of professional recruiters. The experiment dataset will if sufficiently large, be split into a training set and a set reserved for absolute final validation only. During the spring semester an AI based recruitment module will be designed, implemented and evaluated. The evaluation process will utilize the experiment to test our hypotheses.
Participants
We have Erik Espenakk and Magnus Johan Knalstad as researchers with Professor Anders Kofod-Petersen as the main supervisor. We have also coordinated with Fredrik Kristian Knalstad, CEO of Ingraphic AS. Ingraphic AS has contributed with ideas and a network of business contacts.
In our study we use a dataset from linkedin the candidates in this dataset can be considered as anonymous participants. We will also contract some recruiters from selected businesses and have them rank candidates from our dataset. These recruiters can also be viewed as anonymous participants of the research.
Paradigm
We will apply a positivist approach to the research, more precisely a scientific method. As explained in the book the scientific method seeks to find all the regular laws or patterns in our universe. Our research questions are constructed to seek answers and provide conclusions. Therefore, the “Positivism”-paradigm is found suitable for the purpose.
Presentation
The presentation from both phases will pose a result in two different documents, a research project and a master thesis. Our research will also be presented for the company and evaluated for their purposes.