Statistical Learning

Credits
6
Types
Compulsory
Requirements
This subject has not requirements , but it has got previous capacities
Department
UB;UAB
This course offers an in-depth exploration of statistical learning theory and advanced data analysis techniques. Students will develop both theoretical understanding and practical expertise in handling complex biological and health-related datasets.
The curriculum begins with the foundations of statistical learning, covering core problems such as classification, regression, and clustering, as well as essential concepts like loss functions, model complexity, regularization, and figures of merit from signal detection theory. Building on this foundation, students will master preprocessing methods necessary for analyzing real-world data from sources such as chromatography-mass spectrometry and microarrays.
A strong emphasis is placed on dimensionality reduction, including both feature selection and extraction, to address the challenges of high-dimensional biological data. Students will engage with a comprehensive suite of machine learning algorithms, from basic classifiers and clustering techniques to advanced methods such as support vector machines, decision trees, random forests, and neural network architectures.
The course integrates robust validation strategies to ensure reliable model assessment and interpretation.

Teachers

Person in charge

Others

Weekly hours

Theory
2
Problems
0
Laboratory
2
Guided learning
0
Autonomous learning
6

Competences

Learning Outcomes

Knowledge

  • K2 - Identify mathematical models and statistical and computational methods that allow for solving problems in the fields of molecular biology, genomics, medical research, and population genetics.
  • K3 - Identify the mathematical foundations, computational theories, algorithmic schemes and information organization principles applicable to the modeling of biological systems and to the efficient solution of bioinformatics problems through the design of computational tools.
  • K4 - Integrate the concepts offered by the most widely used programming languages in the field of Life Sciences to model and optimize data structures and build efficient algorithms, relating them to each other and to their application cases.
  • K5 - Identify the nature of the biological variables that need to be analyzed, as well as the mathematical models, algorithms, and statistical tests appropriate to develop and evaluate statistical analyses and computational tools.

    • Skills

  • S2 - Computationally analyze DNA, RNA and protein sequences, including comparative genome analyses, using computation, mathematics and statistics as basic tools of bioinformatics.
  • S3 - Solve problems in the fields of molecular biology, genomics, medical research and population genetics by applying statistical and computational methods and mathematical models.
  • S4 - Develop specific tools that enable solving problems on the interpretation of biological and biomedical data, including complex visualizations.
  • S8 - Make decisions, and defend them with arguments, in the resolution of problems in the areas of biology, as well as, within the appropriate fields, health sciences, computer sciences and experimental sciences.

    • Competences

  • C3 - Communicate orally and in writing with others in the English language about learning, thinking and decision making outcomes.
  • C6 - Detect deficiencies in the own knowledge and overcome them through critical reflection and the choice of the best action to expand this knowledge.

    • Objectives

    1. Implement correct data partition schemes for training, optimization and performance assessment in predictive modeling.
      Related competences: K2, K3, K4, S3, S4,
    2. Select proper data preprocessing techniques before model building
      Related competences: K2, K3, K4,
    3. Perform dimensionality reduction using both feature selection and extraction methods.
      Related competences: K2, K3, K4, K5, S3, S4, C3, C6,
    4. Critically assess model performance using appropriate validation techniques.
      Related competences: K2, K3, K4, S3, S8, C3, C6,
    5. Apply advanced machine learning and signal processing methods to real-world bioinformatics and health data challenges.
      Related competences: K2, K3, K4, K5, S2, S3, S4, S8, C3, C6,
    6. To write a lab report in formal language, good structure and good quality visuals
      Related competences: C3,
    7. Orally defend a teamwork regarding a machine learning analysis of a dataset. Produce good-quality slides and structure the presentation to provide a clear message to the audience. Answer technical questions with proficiency.
      Related competences: S8, C3,
    8. Understand technical literature in the area of statistical learning for health. Identify key concepts and identify ideas that required deeper analysis.
      Related competences: K2, K3, C6,

    Contents

    1. Introduction to Statistical Learning: Basic Concepts and Examples
      Motivation and basic concepts: Application examples. Tools
    2. Introduction to Statistical Learning (II)
      Figures of merit. Basic Classifiers. Overfitting i complexity control. Dimensionality reduction. Regularization.
    3. Data preprocessing: from raw data to features.
      Examples in spectrometry. Noise reduction, baseline correction, peak detection and integration, alignment, non-linear transformations, scaling and normalization techniques.
    4. Dimensionality reduction: Feature extraction
      The curse of dimensionality. Principal Component Analysis. Linear Discriminant Analysis.
    5. Dimensionality Reduction: Feature Selection
      The importance of data partition. Univariate approaches. Mutivariate approches: Filters, Wrappers, Sequential Searches, Genetic Algorithms. Feature Rankings and Recursive Feature Elimination.
    6. Clustering
      K-means, Hierarchical clustering, Gaussian Mixture Models, Parzen Windows
    7. Basic classifiers
      Bayes Theorem. Linear and Quadratic Discriminant Classifiers. Naive Bayes. Partial Least Squares Discriminant Analysis.
    8. Model validation and cross-validation
      Validation levels and purpose. Estratificació. Internal/External validation. Hold-out, Leave-one-out, k-fold, random subsampling, Bootstrap
    9. Advanced classifiers
      Support Vector Machines, Decision Trees, Random Forest. XGBoost
    10. Multilinear Regression
      Overview of univariable linear regression. Multilinear Regression. The condition number. Ridge Regression. LASSO. Subset selection.
    11. Advanced Regression
      Neural Networks, The perceptron. The multilayer perceptron. Gradient descent techniques. Deep Learning. Support Vector Regression.

    Activities

    Activity Evaluation act



    Computational lab


    Objectives: 1 2 3 4 5 6
    Theory
    0h
    Problems
    0h
    Laboratory
    30h
    Guided learning
    0h
    Autonomous learning
    22.5h

    Small Project


    Objectives: 1 2 3 4 5 7
    Theory
    0h
    Problems
    0h
    Laboratory
    0h
    Guided learning
    0h
    Autonomous learning
    30h


    mid term exam



    Theory
    2h
    Problems
    0h
    Laboratory
    0h
    Guided learning
    0h
    Autonomous learning
    0h

    Teaching methodology

    The teaching methodology combines lectures with computational laboratories. Additionally, students in groups will have to analyze a set of data and present their analysis orally.

    Evaluation methodology

    The evaluation of the course will take into account the partial exam (P) , the final exam (F), the lab reports (LR) , lab questionnaires (LQ), Reading questionnaire (RQ) the computational homework (H) and the Small Project (SP). They will be combined according to the formula.
    Grade= 0.2*P+0.2*F+0.2*SP+0.05*RQ+0.1*H+0.15*LR+0.1*LQ

    In the case of repeating students, in no case will activities carried out in previous years be taken into account.

    Students who fail the subject may take the reassessment exam; in this case, the grade of this exam, E, will replace the grades P and F so that the final grade will be 0.4*E+0.2*SP+0.05*RQ+0.1*H+0.15*LR+0.1*LQ

    Bibliography

    Basic

    Previous capacities

    Programming in R. Biostatistics. Algebra.