catplot

load_dataset

df.head()

df.heag()

1. Use machine learning to better predict risk

1. Create an algorithm that consolidates all of your Excel

   spreadsheets into one data lake.

1. Create an artificial neural network that would host the company directory.

1. Use machine learning and big data to research salary requirements.

unsupervised machine learning

reinforcement learning

supervised machine learning

binary classification

Linear Regression

PCA

Decision Tree

Naive Bayesian

The algorithms are better at seeing patterns in the data

The algorithms are typically run more powerful servers.

The algorithms can run on unstructured data.

Machine learning servers can host larger databases.

1. identify a smaller number of uncorrelated variables, also called "principal components", from a large set of data

1. Result of the multiplication matrix

1. None of the above

1. identify a large number of correlated variables, also called "main components", from a smaller set of data



1. Unstructured data is always text.

1. Structured data is much more popular.

1. Unstructured data is much easier to store.

1. Structured data has clearly defined data types

1. If the data is high quality, the algorithms will be easier to develop.

1. If the data is low quality, you will get inaccurate results.

1. It will take too long for programmers to scrub poor data.

1. Low-quality data requires much more processing power than high-quality data.

1. Matrix of principal components

1. Set of all eigen vectors for the projection space

1. Result of the multiplication matrix

1. None of the above

1. Bias is when your values are always off by the same percentage.

2. Bias is when your predicted values are scattered

Bias is the gap between your predicted value and the outcome.

Bias is when your data is wrong for different reasons.

False

True

All of the above

Assign a unique category to missing values

Drop missing rows or columns

Replace missing values with mean/median/mode

Eval-Learn

Learn

All of the above

Scikit-learn

Random Forest

Classification

Decision Tree

Regression

X=df

2. X = df.head

2. y = np.array(y)

2. X = np.array(X)

1. K-nearest neighbor

1. learning algorithm are you using

1. binary classification

1. K-means lustering

Include training email data from new employees.

2. Include training email data from employees who write the majority of internal emails.

2. Include training email data from seasoned employees.

2. nclude training email data from all employees.

1. None of the above

1. processed and other linear of data

1. unprocessed and unstructured array of data

1. processed and structured array of data

algorithm allows us to separate data that cannot be separated by a straight line

2. the algorithm does not allow dividing data that cannot be divided in a straight line

2. All the above

he algorithm does not allow you to separate multiple data that cannot be separated by a line

structured data

availability

Statistics

algorithms

1. is provided with labels, and a model learns use of the dataset

1. is provided with labels, and a model learns use properties of the structure of the dataset

All of the above

is provided without labels, and a model learns useful properties of the structure of the dataset

it is an unsupervised preprocessing task that is performed before any ML algorithm is applied

2. an unsupervised algorithm that simultaneously performs dimensionality reduction and clustering

2. All the above

2. based on “orthogonal linear transformation”

1. SHap Additive

1. SHap Additive Predictions

1. SHapley Additive exPlanations

1. All of the above

In K-means clustering, the initial centroids are sometimes randomly selected.

2. To be accurate, you want your centroids outside of the cluster

2. K-means clustering is often used in supervised machine learning.

2. The number of clusters are always randomly selected.

1. to convert strong learners to weak learners

1. All of the above

1. to convert weak learners to strong learners

1. It’s meta-algorithm, which takes M subsamples

iris_df.isnull().sum()

iris_df.sum()

iris.sum()

iris_df.null().sum()

1. a dataset is provided without labels, and a model learns useful properties of the structure of the dataset.

1. where you have input variables (x) and an output variable (Y) and you use an algorithm to learn the mapping function from the input to the output Y = f(X) .

1. where you have other variables (x) and an next variable (Y) and you use an algorithm to learn the mapping data from the other to next Y = f(X) .

1. All of the above

Is representative of the dataset as a whole

Large enough to yield meaningful results

Drop missing

Drop missing rows or columns

1. pivot table

1. K-means cluster diagram

1. multiclass classification diagram

1. linear regression and scatter plots

1. stochastic neighbor

1. high-dimensional space

1. t-distributed stochastic neighbor embedding

1. All the above

Classification

Reinforcement

Supervised

Unsupervised

1. Factor

1. “clusters” and “noise”.

1. Underfitting

1. All the above

1. wcss1 = [3]

1. wcss = []

1. wcss2 = [2]

1. wcss1= [0]

1. to convert weak learners to strong learners

1. All of the above

1. to convert strong learners to weak learners

1. It’s meta-algorithm, which takes M subsamples

1. learning algorithm that predicts a unreal-valued output based on output values

1. unsupervised learning algorithm that predicts a real-valued output based on output values

1. All the above

1. supervised learning algorithm that predicts a real-valued output based on input values

Dlearn

Sklearn

Bklearn

Vlearn

1. underfitting

1. all of the above

1. overfitting

1. good fitting

1. Machine learning algorithms are powerful enough to eliminate bias from the data.

1. All human-created data is biased, and data scientists need to account for that

1. There is no way to identify bias in the data.

1. Machine learning algorithms are based on math and statistics, and so by definition will be unbiased.

1. Unstructured data is much easier to store.

1. Unstructured data is always text.

1. Structured data has clearly defined data types

1. Structured data is much more popular.

1. Unsupervised non-linear dimensionality reduction technique for data exploration and visualizing  high-dimensional data

1. supervised non-linear dimensionality reduction technique for data exploration and visualizing lown- dimensional data

1. Multi-output problems

1. All the above

1. You use several machine learning algorithms to boost your results

1. The predictions of one model become the inputs another

1. You stack your training set and testing set together

1. You use different versions of machine learning algorithms

data

programs

rules

patterns

making the decision boundaries more regula

normalizing the data points so they can be compared with each other

finding the slope of the line that represents the model

using a straight line model to make predictions based on training data

1. is provided with labels, and a model learns use of the dataset

1. is provided with labels, and a model learns use properties of the structure of the dataset

1. is provided without labels, and a model learns useful properties of the structure of the dataset

1. All of the above

1. Leaf/Terminal Node

1. Parent and Child Node

1. Information gain and Gini index

1. All of the above

1. Data

1. Networking

1. statistical techniques

1. Processors

Is representative of the dataset as a whole

Large enough to yield meaningful results

A and B

Drop missing rows or columns

Replace missing values with mean/median/mode

Assign a unique category to missing values

Drop missing rows or columns

All of the above

identify a smaller number of uncorrelated variables, also called "principal components", from a large set of data

identify a large number of correlated variables, also called "main сomponents", from a smaller set of data

Result of the multiplication matrix

None of the above

1. min_samples_leaf

1. min_impurity_decrease

1. max_depth

1. min_samples_split

when you create an ensemble of your training and test data set

when you create an ensemble of different servers to run the algorithms

when you find the one best algorithm for your ensemble

when you use several ensembles of machine learning algorithms

by Leo Breiman to refer to Decision Tree algorithms that can be used for classification and regression modeling problems.

by Neo to refer to Decision Tree algorithms that can't be used for regression modeling problem

by Deo to refer to Decision Tree algorithms that can't be used for classification modeling problems

None of the above

tree nodes

predictors

root nodes

deciders

All of the above

s provided with labels, and a model learns use of the dataset.

is provided without labels, and a model learns useful properties of the structure of the dataset.

is provided with labels, and a model learns use properties of the structure of the dataset.

Density-Based Clustering Spatial Of Applications With Cluster

Density-Based Spatial Clustering Of Applications With Noise

Density-Base Clustering Applications Of Spatial With Data

Density-Base Cluster Application Of Spatial With algorithm

All of the above

Object, Characteristics of the object

especially analysis

statistics and probability

"boolean aggregation".

bootstrap aggregating".

bootstep aggregating".

All of the above

Machine learning algorithms are powerful enough to eliminate bias from the data.

All human-created data is biased, and data scientists need to account for that.

here is no way to identify bias in the data.

Machine learning algorithms are based on math and statistics, and so by definition will be unbiased.