Machine Learning in Bioinformatics 2017
Almost two thirds of estimated 56 million deaths per year worldwide are caused by noncommunicable diseases such as cardiovascular disease, various types of cancer, diabetes and chronic lung conditions (http://www.globalissues.org/issue/587/health-issues). In order to develop an effective treatment, we need to profoundly understand internal mechanisms and sources of these diseases. In this course, we will learn how bioinformaticians use machine learning to help biologists understand, diagnose and develop drugs against human disorders. We will use learning through doing approach by practicing new skills on a real-life dataset.
- Lecture 1: Introduction
- Brief introduction to bioinformatics
- Machine learning recap
- Applications of machine learning in bioinformatics
- The data mining approach
- Major challenges
- Lecture 2: Data exploration and unsupervised learning
- How biological data looks like?
- Descriptive statistics and visualization
- Handling noise and missing data
- Feature selection and clustering
- Lecture 3: Supervised learning
- Linear model and decision tree algorithms
- Overfitting and cross-validation
- Parameter interpretability in biological studies
- Other major classification algorithms
- Model performance estimation
Вимоги до попередніх знань
Participants should be familiar with R programming language, basic calculus and linear algebra. No biological or machine learning background is required.
I am a Junior Researcher and a PhD candidate at the University of Tartu. My current research is about applying machine learning and data mining methods to biological data. I am building an automatic tool for analysing protein microarray experiments in immunological studies for my PhD thesis. I am actively involved into several projects where we use deep learning on genomic data and microscopy images. Also have experience teaching various machine learning related subjects both at the University of Tartu and as an invited lecturer in companies. I am a certified trainer in Data and Software Carpentry organisations that organise and carry out trainings for scientists in the core data science skills around the world.
Галузі професійних інтересів: Data Mining, Machine Learning, Bioinformatics, Image Recognition, Deep Learning, Advanced Algorithms
Контакти: [email protected]