Analysis Basics

Question 1

What do you use to visualize the distribution or spread of a variable?

Answer 1

1. Histogram
2. Box plot

Question 2

What do you do to understand the distribution?

Answer 2

Examine the “measured of central tendency”. This refers to describing the “middle” of the data by getting the mean, median, and mode.

Question 3

A simple average based on adding together all of the values in the sample set and then dividing the total by the number of samples.

Answer 3

Mean

Question 4

The value in the middle of the range of all of the sample values.

Answer 4

Median

Question 5

The most commonly occurring value in the sample set

Answer 5

Mode

Question 6

This refers to a tie for the most common value.

Answer 6

Bimodal or Multimodal

Question 7

It refers to the data that we have on hand.

Answer 7

Samples

Question 8

It refers to all the data that we can collect.

Answer 8

Population

Question 9

Which function is used to estimate the distribution of a variable for the full population?

Answer 9

Probability Density Function (PDF)

Question 10

What type of distribution has the mean and mode at the center and symmetric tail?

Answer 10

Normal Distribution

Question 11

What type of distribution has the “bell shape” characteristic?

Answer 11

Normal Distribution

Question 12

It refers to a tendency to select certain types of values more frequently than others, in a way that misrepresents the underlying population, or ‘real world’.

Answer 12

Bias

Question 13

What are the things to remember when examining real world data?

Answer 13

1. Check for missing values and badly recorded data
2. Consider removal of obvious outliers
3. Consider what real-world factors might affect your analysis and consider if your dataset size is large enough to handle this
4. Check for biased raw data and consider your options to fix this, if found

Question 14

It is a value that lies significantly outside the range of the rest of the distribution.

Answer 14

Outlier

Question 15

Which type of distribution has the mass of the data on the left side of the distribution, creating a long tail to the right because of the values at the extreme high end, which pull the mean to the right.

Answer 15

Right skewed

Question 16

How do you measure variability (variance) in the data?

Answer 16

1. Range
2. Variance
3. Standard Deviation

Question 17

This refers to the difference between the maximum and minimum. There’s no built-in function for this, but it’s easy to calculate using the min and max functions.

Answer 17

Range

Question 18

This refers to the average of the squared difference from the mean. You can use the built-in var function to find this.

Answer 18

Variance

Question 19

This refers to the square root of the variance. You can use the built-in std function to find this.

Answer 19

Standard Deviation

Question 20

It is a built-in method of the DataFrame object that returns the main descriptive statistics for all numeric columns.

Answer 20

df.describe()

Question 21

When comparing numeric variables, how do you deal with numeric data in different scales?

Answer 21

Normalize the data

Question 22

It is a technique that distributes the values proportionally on a scale of 0 to 1.

Answer 22

MinMax scaling

Question 23

This indicates the strength of the relationship between variables.

Answer 23

Correlation

Values above 0 indicate a positive correlation (high values of one variable tend to coincide with high values of the other), while values below 0 indicate a negative correlation (high values of one variable tend to coincide with low values of the other).

Question 24

What do you use to visualize the correlation between two numeric variables?

Answer 24

1. Scatter plot
   2.

Question 25

It is added to a scatter plot that shows the general trend in the data.

Answer 25

Regression line (line of best fit)

Question 26

What is the slope-intercept form of a linear equation?

Answer 26

y = mx + b Where: - y and x are the coordinate variables - m is the slope of the line - b is the y-intercept (where the line goes through the axis)

Question 27

It is the line that gives us the lowest value for the sum of the squared errors

Answer 27

Least Squares Regression

Question 28

This returns (among other things) the coefficients you need for the slope equation: slope (m) and intercept (b) based on a given pair of variable samples you want to compare.

Answer 28

linregress method

Question 29

It is the process of taking a set of sample data that includes one or more features (in this case, the number of hours studied) and a known label value (in this case, the grade achieved) and use the sample data to derive a function that calculates predicted label values for any given set of features.

Answer 29

Machine Learning

Question 30

It works by establishing a relationship between variables in the data that represents characteristics known as the features of the thing being observed and the variable that we’re trying to predict known as the label.

Answer 30

Regression

Question 31

It is an example of a supervised machine learning technique in which you train a model to predict a numeric label based on an item's features.

Answer 31

Regression

Question 32

It is the difference between a predicted label value and the actual label value as a measure of error.

Answer 32

Residuals

Question 33

What are the kinds of Linear Regression algorithms?

Answer 33

1. Least Squares 2. Lasso 3. Ridge

Question 34

What are the kinds of Regression algorithms?

Answer 34

1. Linear Regression 2. Tree-based 3. Ensemble

Question 35

These are algorithms that build a decision tree to reach a prediction.

Answer 35

Tree-based Regression Algorithm

Question 36

These are algorithms that combine the outputs of multiple base algorithms to improve generalizability.

Answer 36

Ensemble Algorithm

Question 37

This algorithm work by combining multiple base estimators to produce an optimal model, either by applying an aggregate function to a collection of base models (sometimes referred to a bagging) or by building a sequence of models that build on one another to improve predictive performance (referred to as boosting).

Answer 37

Ensemble Algorithm

Question 38

This algorithm work by using a tree-based approach in which the features in the dataset are examined in a series of evaluations, each of which results in a branch in a decision tree based on the feature value. At the end of each series of branches are leaf-nodes with the predicted label value based on the feature values.

Answer 38

Decision Tree Algorithm

Question 39

This optimization method works by applying an aggregate function to a collection of base models.

Answer 39

Bagging

Question 40

This optimization method works by building a sequence of models that build on one another to improve predictive performance

Answer 40

Boosting

Question 41

This boosting algorithm is similar to a Random Forest algorithm which builds multiple trees; but instead of building them all independently and taking the average result, each tree is built on the outputs of the previous one in an attempt to incrementally reduce the loss (error) in the model.

Answer 41

Gradient Boosting

Question 42

This bagging algorithm applies an averaging function to multiple Decision Tree models for a better overall model.

Answer 42

Random Forest

Question 43

This refers to changes you make to your data before it's passed to the model.

Answer 43

Preprocessing

Question 44

This refers to the values that you specify to affect the behavior of a training algorithm are more correctly.

Answer 44

Hyperparameters

Question 45

This algorithm is an ensemble that combines multiple decision trees to create an overall predictive model.

Answer 45

Gradient Boosting Regressor

Question 46

Provide preprocessing transformations to get your data for modeling.

Answer 46

1. Scaling numeric features 2. Encoding categorical variables

Question 47

Provide techniques to encode categorical variables.

Answer 47

1. Ordinal encoding 2. One-hot encoding

Question 48

How do you save a model?

Answer 48

joblib.dump(model, filename)

Question 49

How do you load a model

Answer 49

model = joblib.load(filename)

Question 50

How do you use a model to generate predictions?

Answer 50

model.predict

Question 51

You have created a model object using the scikit-learn LinearRegression class. What should you do to train the model?

Answer 51

Call the fit() method of the model object, specifying the training feature and label arrays

Question 52

It is a measure of how much of the variance the model can explain.

Answer 52

R squared metri

Question 53

It works by establishing a relationship between variables in the data that represent characteristics—known as the features—of the thing being observed, and the variable we're trying to predict—known as the label.

Answer 53

Regression

Question 54

This metric for measuring loss in a regression squares the individual residuals, sum the squares, and calculate the mean. Squaring the residuals has the effect of basing the calculation on absolute values (ignoring whether the difference is negative or positive) and giving more weight to larger differences.

Answer 54

Mean Squared Error or MSE

Question 55

This metric for measuring loss in a regression is calculated by getting the square root of the MSE. This is to express the loss in the same unit of measurement as the predicted label value itself.

Answer 55

Root Mean Squared Error or RMSE

Question 56

This metric for measuring loss in a regression is also known as coefficient of determination.

Answer 56

R squared

Question 57

This metric for measuring loss in a regression is the correlation between x and y squared. This produces a value between 0 and 1 that measures the amount of variance that can be explained by the model. Generally, the closer this value is to 1, the better the model predicts.

Answer 57

R squared

Question 58

It is a container that holds related resources for an Azure solution

Answer 58

Resource Group

Question 59

What are the 4 kinds of compute resource?

Answer 59

1. Compute instances 2. Compute clusters 3. Inference clusters 4. Attached compute

Question 60

These are development workstations that data scientists can use to work with data and models.

Answer 60

Compute instances

Question 61

These are scalable clusters of virtual machines for on-demand processing of experiment code.

Answer 61

Compute clusters

Question 62

These are deployment targets for predictive services that used your trained models.

Answer 62

Inference clusters

Question 63

These are links to Azure compute resources, such as Virtual Machines or Azure Databricks clusters.

Answer 63

Attached compute

Question 64

It is the variance between predicted and true values that cannot be explained by the model.

Answer 64

Residuals

Question 65

It is an example of a supervised machine learning technique in which you train a model to predict a numeric label based on an item's features.

Answer 65

Regression

Question 66

It is a cloud-based platform for building and operating machine learning solutions in Azure.

Answer 66

Microsoft Azure Machine Learning

Question 67

It refers to the average difference between predicted values and true values. The lower this value is, the better the model is predicting.

Answer 67

Mean Absolute Error (MAE)

Question 68

It refers to the square root of the mean squared difference between predicted and true values. When compared to the MAE (above), a larger difference indicates greater variance in the individual errors (for example, with some errors being very small, while others are large).

Answer 68

Root. Mean Squared Error (RMSE)

Question 69

It refers to a relative metric between 0 and 1 based on the square of the differences between predicted and true values. The closer to 0 this metric is, the better the model is performing. Because this metric is relative, it can be used to compare models where the labels are in different units.

Answer 69

Relative Squared Error (RSE)

Question 70

It refers to a relative metric between 0 and 1 based on the absolute differences between predicted and true values. The closer to 0 this metric is, the better the model is performing. Like RSE, this metric can be used to compare models where the labels are in different units.

Answer 70

Relative Absolute Error (RAE)

Question 71

It summarizes how much of the variance between predicted and true values is explained by the model. The closer to 1 this value is, the better the model is performing

Answer 71

Coefficient of Determination