ML Flashcards
List the 04 main types of learning.
- Unsupervised
- Supervised
- Reinforcement
- Semi-supervised
What is machine learning?
The study of algorithms that can learn from data and give predictions on data. (mostly text and numerical data)
What is the difference between classification and regression?
- classification -> when the label is discrete
- regression -> when continuous
List 04 examples for regression models.
- Linear regression
- polynomial regression
- lasso regression
- ridge regression
List 06 popular classifiers.
- Neural networks
- SVM
- Decision tree based methods
- Probabilistic graphical methods(naive bayes)
- Nearest neighbor classifiers
- Meta learning methods
What are the 08 steps of a workflow of an ML project?
- Understanding the business, prior knowledge and goals
- Data gathering
- Data preprocessing and EDA
- Feature engineering
- Find the best learning models/ algorithms
- Evaluating the model and hyperparameter tuning
- Consolidating and deploying the model
- Customer acceptance and consumption
List 06 things considered in EDA
- Missing data
- Noisy data
- Correlated data
- Inconsistent data
- Conversion of data
- Outlier detection
List 05 domain specific techniques that may be used in feature engineering and selection
- Image processing
- Language processing
- Signal processing
- Mathematics
- Statistics
What are hyperparameters?
The values of the parameters that will affect the accuracy scores the most.
List 06 performance measures used in evaluation and hyperparameter tuning
- ROC
- F1
- Precision
- Accuracy
- Confusion matrix
- AIC
List 03 methods to evaluate the accuracy of a model
- performance measures
- cross validation
- hold out
What is Deep Learning?
Deep learning is the way of implementing ML via artificial neural networks, which are algorithms that loosely mimic the human brain’s structure and function.