Titulo Estágio
ML4PM - Machine Learning for Project Management
Áreas de especialidade
Engenharia de Software
Engenharia de Software
Local do Estágio
Porto or Remote Work
Enquadramento
Summary
The main idea is the adaptation of the current Project Control to then create and deploy a Machine Learning (ML) model based on project management information.
Software that manages teams, projects and budget is the backbone of any company’s business nowadays not just the simple execution of well defined tasks.
At Fraunhofer AICOS, all projects are managed using Project Control, an internally developed platform that provides assistance in managing projects, human resources and budgeting. The purpose of this master thesis is to apply machine learning techniques to Project Control’s database and to investigate which recommendations can be provided to Project Managers to assist in the project management task. Project Control is currently supported by a relational database, however, to be able to apply machine learning techniques some adaptations are required.
In the context on this dissertation the student will need to adapt the current data structure, to evaluate the data and to detect and create new features that characterise (patterns) the projects / teams. With this information will be possible to infer relevant information like, e.g., which tasks usually cause losses and/or are most commonly delayed, or what expenses are usually not spent in time, estimations of workload on tasks, proper allocation of human resources to tasks regarding their competences and task importance, among many other possible conclusions
Objetivo
Objectives and expected results
The main goal is to create a model to that infers relevant information regarding a project’s health and performance. To that it is necessary to adapt the current Project Control Database to support ML techniques.
Innovative Aspects
Automatic detection of patterns within a company to support project management.
Plano de Trabalhos - Semestre 1
Workplan
1. Literature Review:
- Applying Machine Learning techniques on structured data: Research in methods of how to prepare structured data from relational databases to apply machine learning techniques in a efficient and correct way.
- Supervised, Unsupervised and Deep Learning: Research in Machine learning algorithms in the area of Project management. It is expected to weight the pros and cons of supervised, unsupervised and deep learning techniques in the context of Project Management.
- Analysis of important features to Project Management: Analysis of the available data and evaluate the features that most influence the decisions a project manager must do regarding an active project.
Plano de Trabalhos - Semestre 2
Workplan (cont.)
2. System development:
- Migration from Relation to Non-Relational Database: Most Machine Learning Techniques require that the input data for learning must be flatten. To automatize this process the student will migrate the existent databases to non-relational databases that ease the process of flattening. This task also includes the preparation of the input data (cleaning data).
- Creation and tuning of the Model: Select, train and tune several models for the specific problems of project management and compare their efficacy so a final model is selected.
3. Evaluation:
- Testing of the model with real users: Prepare a testing protocol for real users (recruited from Fraunhofer) and extract measures related with the information created and/or how it is displayed.
4. Documentation.
- Prepare the documentation of the models tested and created, i.e. information used, models trained, results attained and the necessary steps to setup the projects created.
Condições
Candidate Profile
- Knowledge in Machine Learning.
- Knowledge in OOP (Python and/or Java/Kotlin).
- Good communication skills both written and oral.
Observações
References
- P. N. Robillard, “The Role of Knowledge in Software Development,” Communications of the ACM, vol. 42, no. 1, pp. 87–92, 1999
- https://www.atlassian.com/blog/software-teams/3-ways-ai-will-change-project-management-better
- https://www.atlassian.com/blog/software-teams/3-ways-ai-will-change-project-management-better
- Perini, A., Susi, A., Avesani, P., A machine learning approach to software requirements prioritization (2013), IEEE Transactions on Software Engineering 39(4), pp. 445-461.
- Fürnkranz, Johannes, Hüllermeier, Eyke (Eds.), Preference Learning (2010), Springer. , Aggarwal, Charu C. (2016). Recommender Systems: The Textbook. Springer. ISBN 9783319296579.
- João Mendes Moreira, André de Carvalho, Tomás Horvath, A general introduction to data analytics (2018)
- Anthony S. Tay, Kenneth F. Wallis, Density Forecasting: A Survey (2000), Journal of Forecasting, 19, pp. 235-254. )
- Lukasz Kurgan and Petr Musilek. A survey of knowledge discovery and data mining process models. The Knowledge Engineering Review, 21(1):1–24, 2006.
- Hugh Watson. The crisp-dm model: The new blueprint for data mining. Journal of Data Warehousing, 5(4):13–22, 2000.
- Wiley. , Oliveira, A.L.I., Braga, P.L., Lima, R.M.F., Cornélio, M.L., GA-based method for feature selection and parameters optimization for machine learning regression applied to software effort estimation (2010), Information and Software Technology 52(11), pp. 1155-1166.
Orientador
Rui Nuno Pires Sarmento e Castro
rui.castro@fraunhofer.pt 📩