UniTO/anno3/apprendimento_automatico/programma.org
2024-10-29 09:11:05 +01:00

60 lines
2.6 KiB
Org Mode
Executable file

* Programma AA 2018/2019 - 6 crediti
- Introduction OK
- Differences between tasks and models OK
- Binary classification and transformation of a binary classification model into a multiple class model OK
- Concept learning by means of logical formulas, OK
- Tree models and their purposes, OK
- Rule models, OK
- Subgroup discovery OK
- Linear models (least squares, regression), perceptron, OK
- Support Vector Machines and Kernel methods. OK
- Bias-variance decomposition ?? OK ?? capitoli prima
- The problem of overfitting. ?? OK ?? capitoli prima
- Models based on distance, such as k-nearest neighbors OK
- K-means clustering OK
- Hierarchical clustering, OK
- Density based clustering (DBSCAN) OK
- Validation and verification of the results on data (5-cross validation? Fuffa) OK
- Ensemble learning (not in official program) OK
Experiments with a real system on data sets: Scikit Learn
/Note: chap_12 experiments not asked at the exam but might be useful for scikit/
* Diff 9 / 6 crediti
- Distinctions between types of features, the allowed operations and statistical descriptions, MAGARI LO FAMO LO STESSO EH??
- Transformation among features (normalization, discretization, calibration, etc). MAGARI LO FAMO LO STESSO EH??
- Probabilistic models (such as maximum likelihood estimation, logistic regression, Bayes models and naive Bayes, Expectation-Maximization). NONONO.jpg
- Evaluation of the models and of the statistical significance tests on the results. PFFFF.jpg
* Programma AA 2018/2019 - 9 crediti
- Introduction
- Differences between tasks and models
- Binary classification and transformation of a binary classification model into a multiple class model,
- Concept learning by means of logical formulas,
- Tree models and their purposes,
- Rule models,
- Subgroup discovery,
- Linear models (least squares, regression), perceptron,
- Support Vector Machines and Kernel methods.
- Bias-variance decomposition
- The problem of overfitting.
- Models based on distance, such as k-nearest neighbors
- K-means clustering
- Hierarchical clustering,
- Density based clustering.
- Distinctions between types of features, the allowed operations and statistical descriptions,
- Transformation among features (normalization, discretization, calibration, etc).
- Probabilistic models (such as maximum likelihood estimation, logistic regression, Bayes models and naive Bayes, Expectation-Maximization).
- Evaluation of the models and of the statistical significance tests on the results.
- Validation and verification of the results on data
- Experiments with a real system on data sets: Scikit Learn