Machine learning: fundamental tools and practices - course RUB 51,590. from Netology, training 10 months, Date November 30, 2023.

Model building
Learn what the Sklearn library is and how to use it. Learn clustering algorithms and be able to build ensembles of models. Learn to evaluate models and work with overfitting. You will learn how to use GridSearch and RandomizedSearch, Model Specific CV, Out of Bag approach.
• Sklearn library
• Classification algorithms: linear methods, logistic regression and SVM
• Classification algorithms: decision trees
• Regression algorithms: linear and polynomial
• Clustering algorithms
• Ensemble
• Model accuracy assessment, retraining, regularization
• Improving the quality of the model
• Project organization, preparation of research reports
• Laboratory work
• Delivery of the intermediate project

Working with the customer
You will learn to plan the development of data science projects, as well as competently tell customers about research results.
• Project organization
• Preparation of research reports

Recommender systems
In this and the following blocks, you will apply the acquired knowledge in different areas of machine learning. During this block, learn how to build personalized and non-personalized recommendation systems, and how to combine them.
• Introduction and classification of recommender systems
• Content-based recommendations
•Collaborative Filtering
• Non-personalized recommendation systems
• Hybrid algorithms

Computer vision
You will master basic computer vision techniques: feature extraction, image search, segmentation, object detection, and also learn how to build neural networks.
• Search by images
• Image segmentation, object detection
• Application of ultra-precise neural networks for segmentation and detection tasks
• Application of recurrent networks in image processing problems
• Generative Adversarial Networks (GANs)

Natural Language Processing (NLP)
You will master morphological and syntactic analysis, distribution semantics and information retrieval, learn to reduce dimensionality in a vector model, classify, extract information and generate texts.
• Morphological and syntactic analysis
• Methods for reducing dimensionality in a vector model. Information search
• Topic modeling (LSA, LDA, HDP)
• Distributive semantics (word2vec, GloVe, AdaGram)
• Countable language models and probabilistic language models. LSTM. Machine translate
• Text generation (Natural Language Generation)
• Classification problem in AOT

Time series
In this intensive unit, you will learn to identify the origin and structure of a time series, predict future values ​​for effective decision making when building machine learning models. You will understand what is “under the hood” of popular methods and libraries.
• Algorithms for processing time series
• ARIMA and GARCH models
• Markov random processes

Final hackathon
Let’s complete the training by competing with course mates: as part of a mini-team for a limited time and based on datasets of major players market, you will have to solve problems of forecasting sales or optimizing production, using all the knowledge and skills acquired in course. Integration and use of machine learning solutions in business, as a rule, involves team play, so a hackathon is also useful as training the necessary soft skills.

Graduation project
As part of your thesis project, you will build an ML model to solve your current professional problems: this could be a system sales forecasting, object recognition in photos or videos, time series analysis, analysis of large amounts of text, etc. d. If at the moment you do not have ideas for your project (or access to the necessary data), we will offer you a case study in an area of ​​interest to you based on a real dataset of other companies. The thesis is completed independently under the guidance of course experts and allows you to consolidate the entire range of knowledge and skills acquired in the program.