EAE_DataScience_S2

Which is the most typical error when integration data from different sources?

• Merge records that are different

• Create new data

• Keep duplicated records

• Remove outliers

Which is the most used visualization techniques when performing exploratory analysis?

• Scatterplots

• Historgrams

• Box Plots

• All previous ones

What type of correlation can be seen in the following scatterplot?

• No correlation

• Perfect positive correlation

• Perfect negative correlation

• Strong negative correlation

Which step is typical from a Data cleaning phase?

• Feature engineering

• Outliers analysis

• Discretization

• Dimensionality reduction

What type of atribute can contain the following values?

• Numerical

• Ordinal

• Nominal

• None of the previous ones

Which of the following items is not a method to deal with missing values?

• Predict a value

• Filter an instance

• PCA

• Discretize

What is the problem that can arise if you want to take into account all possible features?

• The curse of dimensionality

• Not achieving a convergence

• Creating redundant data

• Losing data

Principal Component Analysis is a supervised Algorithm

• True

• False

If we create an attribute called "Age at the time of purchase", what are we doing?

• Feature reduction

• Discretization

• Feature Engineering

• Convert to categorical

What are we probably doing when we create a “1 to n-1” flag?

• Quantization

• Discretization

• Convert a categorical attribute into numerical

• Convert and ordinal attribute to numerical