Thesis subject

Machine Learning Based Approach for Reducing Malnutrition (MSc)

Level: MSc


Although much progress has been made in toward ending hunger, significant proportions of the population in many countries are still undernourished (1). Although this is particularly problematic in developing countries, developed countries are not exempt; the shift towards nutrient-poor foods and the rise in overweight and obesity in recent years has too left populations developed countries in a state of malnutrition. Malnutrition is also relevant under other circumstances such as the clinical setting, where failure to detect malnutrition can have serious consequences for patient outcomes.

In many of these cases, data is collected that could be used to predict states of malnutrition. Here, novel technologies such as machine learning, a branch of artificial intelligence, can play a role. By being able to predict and categorize malnutrition in these various populations, interventions can be taken to mitigate the negative effects of malnutrition, such as growth stunting and mortality in children, micronutrient deficiency in the developed world, and complications in the clinical setting, amongst others. Despite this potential, a systematic review of machine learning in malnutrition has yet to be performed.

The global prevalence of malnutrition shows that much work is still needed in order to achieve the World Health Organization’s goal of ending all forms of malnutrition by 2030 (2). Machine learning may be able to help in achieving this goal by using data to predict and categorize malnutrition, allowing preventative measures to be implemented by governing health bodies.

Project Aims:

The overall aim of the project will be to provide the first ever systematic literature review investigating machine learning in human malnutrition followed by performing machine learning experiments on a malnutrition-related dataset.

In the first part of the project, a systematic literature review will be performed, where research questions will be established and answered covering themes such as:

  • Machine learning types and tasks in malnutrition;
  • The problem areas tackled within malnutrition;
  • Which algorithms are being implemented, which are performing best, and how are they being graded;
  • Which features (variables) are being used; Which data are being used and what is the availability status of the data

In the second part of the thesis, a primary study will be performed using a data set relevant to malnutrition. The exact data set used will depend on data availability at the time of the project. Depending on time and the capability of the student, multiple primary studies may be performed tackling different problems identified in the systematic review.


Review: Research questions will be devised in order to fully elucidate the role of machine learning in malnutrition. To answer the research questions, a search strategy will be designed which will cover inclusion and exclusion criteria, search terms, and the major search databases to be searched. Abstracts and titles will be screened, articles will be filtered based on the inclusion and exclusion criteria, and the information relevant for answering the research questions will be obtained.

Primary Study: Using a yet to be determined data set, the data science process (data pre-processing and cleaning, data exploration, feature engineering, modelling, evaluation, and visualization) will be applied to predict malnutrition-related outcomes, using answers from the research questions of the review as a guide. If time and student ability permits, various studies covering various topics in malnutrition can be performed.

Our Expectations:

So far, no systematic review of machine learning in malnutrition in the literature has been performed. Thus, this thesis not only represents an excellent opportunity to gain a deep understanding of machine learning, nutrition and malnutrition, and their fusion but also to publish in a scientific journal. To achieve this, we require a driven and self-motivated student with strong capabilities in examining scientific literature and scientific writing. Basic understanding of machine learning and data science concepts is also required.

For the primary study, the student must have at least basic familiarity with data science and machine learning concepts and techniques and know how to apply them to a malnutrition-related data set. Some experience in coding will be required. Python is preferred, although the exact software can be negotiated, especially if the student has a strong preference for other programming languages.

More Information?

For more information on any aspect of the project, feel free to contact:

Bedir Tekinerdogan (

Daniel Kirk (


(1) FAO and IFPRI 2020. Progress towards ending hunger and malnutrition: A cross-country cluster analysis. Rome.

(2) Malnutrition. (2019, November 14). World Health Organization.