"Machine learning on Python" course program
Below you can see a short training program
Or take a look at the detailed program
It has descriptions of the webinars and tasks for each webinar.
Or take a look at the detailed program
It has descriptions of the webinars and tasks for each webinar.
Detailed training program
With descriptions of webinars and tasks
June 15
19:00 Moscow time
19:00 Moscow time
June 15
19:00 Moscow time
19:00 Moscow time
Preparation of data for model building and the simplest algorithms of machine learning
Session topics:
Task: prediction whether or not the animal will be adopted from the DataSet PetFinder (binary classification)
- Primary processing of raw data. The main disadvantages of the source data: "trash", omissions, emissions.
- k- neighbors method
- Processing DataSet Titanic raw data and building a binary classification model
Task: prediction whether or not the animal will be adopted from the DataSet PetFinder (binary classification)
June 22
19:00 Moscow time
19:00 Moscow time
June 22
19:00 Moscow time
19:00 Moscow time
Metrics in binary classification tasks, Bayes classifiers
Session topics:
- Metrics in the tasks of binary classification. Completeness, accuracy, ROC-AUC curve, f-measure
- Bayes Classifiers
- Model building for DataSet Titanic, model quality analysi
June 29
19:00 Moscow time
19:00 Moscow time
June 29
19:00 Moscow time
19:00 Moscow time
Linear models, regression tasks
Session topics:
- Linear regression, least squares method
- The problem of regression and normalization of features
- Model building for DataSet Energy Star Score, model quality analysis
July 6
19:00 Moscow time
19:00 Moscow time
July 6
19:00 Moscow time
19:00 Moscow time
Support vector machine (SVM), prediction of probability of belonging to a class
Session topics:
Task: prediction whether or not an animal will be adopted from the DataSet PetFinder
- Support vector machine (SVM), clear demonstration of the model on the example of DataSet Iris;
- Visual demonstration of the model on the example of DataSeta Iris;
- Determining the probability of class belonging on the example of DataSet Titani.
Task: prediction whether or not an animal will be adopted from the DataSet PetFinder
July 13
19:00 Moscow time
19:00 Moscow time
July 13
19:00 Moscow time
19:00 Moscow time
Decision trees, random forest, gradient boosting.
Session topics:
Task: construction of classic models of machine learning Boston house prices dataset, DataSet PetFinder, comparison of their quality
- Algorithms for building decision trees ID3 and C4.5;
- Random forest in regression and classification tasks;
- Gradient boosting;
- Model building based on decision trees on the example of DataSet Titanic.
Task: construction of classic models of machine learning Boston house prices dataset, DataSet PetFinder, comparison of their quality
July 20
19:00 Moscow time
19:00 Moscow time
July 20
19:00 Moscow time
19:00 Moscow time
Multi-class classification, clustering
Session topics:
- Data clustering algorithms (k-neighbor method, hierarchical methods, t-SNE);
- Models of multiclass classification, evaluation of their parameters on the example of Wine Dataset Recognition.
July 27
19:00 Moscow time
19:00 Moscow time
July 27
19:00 Moscow time
19:00 Moscow time
Solving regression problem
Session topics:
Task: feature selection and feature engineering in predicting house price, Boston house prices dataset
- Feature selection/ feature engineering;
- Building model on the example of DataSet House Price.
Task: feature selection and feature engineering in predicting house price, Boston house prices dataset
August 3
19:00 Moscow time
19:00 Moscow time
August 3
19:00 Moscow time
19:00 Moscow time
NLP task solution
Session topics:
- Regular expressions;
- Methods of numerical representation of text information;
- Model building on the example of DataSet Imdb.
August 10
19:00 Moscow time
19:00 Moscow time
August 10
19:00 Moscow time
19:00 Moscow time
Solutions to the marketing problem of customer churn
Session topics:
Task: dataSet Telecom Churn customer churn prediction
- Statistical analysis of the signs;
- Model building on the example of DataSet Telecom Churn.
Task: dataSet Telecom Churn customer churn prediction
August 17
19:00 Moscow time
19:00 Moscow time
August 17
19:00 Moscow time
19:00 Moscow time
Building a recommendation system
Session topics:
Task: building a movie recommendation system using MovieLens 20M DataSet as an example.
- Approaches to building recommendation systems;
- SVD decomposition, vector representation of objects;
- Model building on the example of MovieLens 20M DataSet.
Task: building a movie recommendation system using MovieLens 20M DataSet as an example.
August 24
19:00 Moscow time
19:00 Moscow time
August 24
19:00 Moscow time
19:00 Moscow time
Identification problem solution
Session topics:
- Methods of image processing;
- Model building on the example of Fashion-MNIST.