Master Data Preprocessing

What is the purpose of exploratory data analysis (EDA)?

To build up a general and detailed picture of the data

What are the types of visualizations used in EDA?

Univariate, bivariate, and multivariate

What are some examples of univariate analysis?

Descriptive statistics, one-sample tests, tests for autocorrelation

What is data imputation?

The process of filling in missing values in real data Signup and view all the answers

Which type of variables can logistic regression be used to predict?

Nominal binary variables Signup and view all the answers

What is the primary use of logistic regression from an econometric perspective?

Inference Signup and view all the answers

Why is interpreting logistic regression results more difficult than interpreting linear regression results?

Because logistic regression results cannot be interpreted directly Signup and view all the answers

What is the focus of logistic regression in this course?

Prediction Signup and view all the answers

What is recommended for learning the principles of logistic regression from an econometric perspective?

Chapter 5.2 of this course Signup and view all the answers

What is the cost function used for logistic regression?

Cross-entropy (log-loss) Signup and view all the answers

What is the purpose of the link function in GLM?

To relate the linear model to the response variable Signup and view all the answers

What is the advantage of using logistic regression over linear regression?

Logistic regression can handle non-linear relationships Signup and view all the answers

What is the difference between binary logistic regression and multinomial logistic regression?

Binary logistic regression can only handle two classes, while multinomial logistic regression can handle more than two classes Signup and view all the answers

What is the purpose of the odds ratio in logistic regression?

To compare the odds of an event occurring in two different groups Signup and view all the answers

What is the purpose of feature engineering during the ETL process?

To transform sets into a form consumable by models Signup and view all the answers

What is the purpose of feature engineering after the ETL process?

To improve the predictive power of the algorithm Signup and view all the answers

What is the purpose of scaling to a range in numeric variable transformations?

To uniformly distribute the feature across a fixed range Signup and view all the answers

What is the purpose of clipping (winsorization) in numeric variable transformations?

To clip the feature if it is greater than max Signup and view all the answers

What is the main challenge in feature engineering after the ETL process?

Capturing non-linear relationships Signup and view all the answers

What is the purpose of multinomial logistic regression?

To classify more than two classes Signup and view all the answers

What is logistic regression?

A selected model representing the entire class of Generalized Linear Models (GLMs) Signup and view all the answers

What is the sigmoid function?

A mathematical function having a characteristic 'S'-shaped curve or sigmoid curve Signup and view all the answers

What are the useful properties of the logistic function?

It maps solution space to probability functions, it is differentiable Signup and view all the answers

What course is used in the text to present the concepts, mathematical foundations, and interpretation of logistic regression?

Machine Learning University (MLU)-Explain course created by Amazon Signup and view all the answers

What is the advantage of machine learning over classical econometrics in terms of feature engineering?

Machine learning algorithms are able to select relevant variables themselves Signup and view all the answers

What are some examples of super powerful encoders mentioned in the text?

Hashing, BaseN, CatBoost Signup and view all the answers

What is a cautionary note given by the author regarding feature engineering in financial problems?

We should not overdo our creativity Signup and view all the answers

What types of interactions can we look for between variables during feature engineering?

Numeric & numeric, categorical & categorical, or numeric & categorical Signup and view all the answers

What are some techniques for dealing with missing values in a dataset?

Fill in the missing values using univariate or multivariate techniques Signup and view all the answers

When should variables/columns with missing values be removed from a dataset?

When the variable has more than 10% missing values Signup and view all the answers

What is feature engineering?

A process for generating new variables Signup and view all the answers

What is one way to fill in missing values for time series variables?

Use the last or next observed value Signup and view all the answers

What is one multivariate technique for dealing with missing values in a dataset?

Use a supervised machine learning algorithm like KNN Signup and view all the answers

What is the purpose of a problem statement worksheet in machine learning projects?

To formalize the definition of the business task Signup and view all the answers

What are the elements of a data preparation process in machine learning projects?

Data selection, data transformation, and data combination Signup and view all the answers

What is the role of consulting firms in machine learning projects?

To formulate a problem statement worksheet Signup and view all the answers

What should be applicable later on the test set in a machine learning project?

Parameters learned on the training set for normalization Signup and view all the answers