Basic knowledege och Machine learning Flashcards
State the three classes of Machine learning, and how it learns in the diffrent classes.
-Supervised Learning: Learning from labeled data
-Unsupervised Learning: Discovers structure in unlabeled data.
-Reinforcement Learning: Optimizes action in unknown environments.
State machine learning applications in Finance
-Banking: Risk estimation, automated trading
-Insurance: Personalized risk assessment, pricing
-General: Fraud detection and automated report analysis.
state the two diffrent Supervised ML Models, and explain the difference.
Decision Trees: Good for smaller datasets and easier.
Neural networks: Suitable for larger datasets but complex
What are the Data types for Supervised ML?
Discrete numerical, continuous numerical, categorical.
what are the assessment performance metrics?
-Accuracy, precision, recall
-Class Imbalance issue: Importance in datasets with skewed class distribution.
What default features does a Risk model have and what are the model types for it?
Risk models includes: Income, Age, employment status..
Logistic regression, Neural network with ReLU
What is the goal of likelihood calculations and model training ?
To minimize average total deviance (That is negative conditional log-likehood)
State three Performance Evaluation techniques.
-Overfitting check, contingency Table Analysis, ROC curve
what are the practical application tools ?
Data simulation: logistic models to create datasets
Google colab, python.
How can you evaluate profit and loss analysis for Non-linear cases?
Examining loan scenarios with varying interest rates
General remarks
Realism and application: Artificial datasets for educational pruposes
Importance of logistic regression
Improvments with feature engineering
