The basics of Machine Learning
Objectives
-
Offer the student a global vision of the following aspects of Machine Learning and Data Science:
-
A large number of examples of success and failure of the most widely used techniques.
-
Intuitive approach to the problem and the explanation of its formalization.
-
Theoretical and practical difficulties of learning and data management.
-
A palette of the most common techniques.
2. To familiarize the student with the mathematical language used in the literature on 3. Machine Learning so that they are able to approach it comfortably.
The ability to successfully and autonomously replicate the theoretical and practical aspects of the simplest algorithms: linear.
Syllabus
1. Some examples of applications
● Classification of Texts and images
● Retail problems
● Optimization
● Prediction of Churn Rate and Medical diagnosis
● Linear and non-linear forecasting
● Artificial intelligence
● Anomaly detection
2. Some technical problems faced by data analysts
● Choosing the right algorithm
● Determination of parameters
● Underfitting vs Overfitting
● The curse of dimension
3. Some simple algorithms intuitively
● Perceptron
● Linear regression
● K-nearest neighbors
● Decision trees
4. Mathematical foundations
● Linear geometry
● Basic probability
● Statistics: law of large numbers
● Convexity
5. Formalism in Machine Learning and case studies
● Classification of false complaints (Revised linear regression).
● Management of financial portfolios (Convex learning)
● Netflix Suggestions (K-nearest neighbors)
● Medical diagnosis (Decision trees)
6. Deep Learning (optional)
● From the perceptron to neural networks
● From linear regression to neural networks
● Advantages: predictability
● Disadvantages: computational cost