Trained Machine Learning Model for Innovation Team Composition Management (MIKiMO)

Project no.: PP-91N/19

Project description:

The project MIKiMO aims to create the machine learning model for innovation team composition management that is trained to predict the team’s potential for innovation based on the intensity of knowledge and diversity in attitudes, and their change in the team composition. A trained machine learning model is converted to a program, that can be used as a product to predict a level of innovation of the team and thus to manage its composition. The project develops interdisciplinary knowledge, where the collected social data serve as the basis for training the machine model, which will be the basis for quick and accurate solutions for the sustainable management of such complex structures as innovation teams. Such model makes it possible to identify weak signals that allow you to anticipate the future performance of an innovation team, and to conduct early management interventions as well as to take important business development solutions (e.g. risk investments). Based on the developed solutions it aims to create a commercialized application for business (start-ups and large innovative organizations).

Project funding:

KTU R&D&I Fund

Project results:

1. Developed an indicator system for determining the innovativeness of a team according to the intensity of knowledge and the diversity of attitudes, and their change in the team composition, and on its basis developed an integrated conceptual model for the prediction of the team’s innovativeness and sustainable development.
2. Trained and validated machine learning model for innovation team composition management.
3. Using a trained machine learning model in creative interdisciplinary innovation teams its predictions for team innovativeness was compared with real results and the guidelines for further model improvement and practical recommendations were provided.

The main result of the project – Web-based team innovativeness prediction engine integrating 6 trained machine learning models with prediction accuracy of 70-90%.

Period of project implementation: 2019-04-01 - 2019-12-31

Vytautė Dlugoborskytė

2019 - 2019

Academic Centre of Economics, Business and Management, School of Economics and Business