Supervised Learning

Question 1

What is supervised learning

Answer 1

Predicts target variables using labeled data (predictor variables)

Question 2

What is unsupervised learning

Answer 2

Uncovers hidden patterns using unlabeled data

Question 3

What is reinforcement learning

Answer 3

Software interacts with environment using system of rewards and punishments to optimize behavior

Question 4

type()

Answer 4

Tells the type of data

Ex: (numpy.ndarray)

Question 5

.shape

Answer 5

Tells the shape of the array or dataset

Ex: (150, 4)

Question 6

.target_names

Answer 6

Shows the dependent variables of the array

Ex: array([‘setosa’, ‘versicor’, ‘virginica’], …)

Question 7

What is k-Nearest Neighbors and how does it work

Answer 7

It predicts label of a data point by looking at ‘k’ closest labeled data points

It takes majority vote

Question 8

.fit()

Answer 8

Uses training data to create a model called “fitting”

Question 9

.predict()

Answer 9

Predicts the labels of new data based on what it learned from the .fit() method

Question 10

What is train_test_split() and what is the function structure

Answer 10

Divides up data into train and test sets to create unbiased prediction models

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3 …)

Question 11

How does larger ‘k’ effect model complexity and smoothness

Answer 11

Less complex model and smoother

Question 12

How does smaller ‘k’ effect model complexity and smoothness

Answer 12

More complex model and less smooth; can lead to overfitting

Question 13

How to read .csv files

Answer 13

pd.read_csv()

Question 14

How to drop an entire column in a dataframe named df

Answer 14

df.drop(‘column_name’, axis=1)

Question 15

How to get values of a column ‘col’ of a dataframe called df

Answer 15

df[‘col’].values

Question 16

Imports

Answer 16

import numpy as np

from sklearn.linear_model import LinearRegression

from sklearn.model_selection import train_test_split

from sklearn.metrics import classification_report

from sklearn.metrics import confusion_matrix

from sklearn.linear_model import LogisticRegression

from sklearn.metrics import roc_curve

from sklearn.model_selection import GridSearchCV

Question 17

How to get accuracy of training model on test model

Answer 17

.score(X_test, y_test)

Question 18

What is cross-validation and how to use it

Answer 18

Cross-validation divides up training data into multiple folds and test data into a single fold to get multiple accuracy scores at once

cross_val_score(regressor, X, y, cv=3)

Regressor can be any chosen such as LinearRegression and there are 3 folds

Question 19

Time how long it takes for programs to run

Answer 19

% timeit before any function

Question 20

What is Lasso regression and how to use it

Answer 20

Can be used to select important features of a dataset by minimizing unimportant variable coefficients to 0

lasso = Lasso(alpha=0.1, normalize=True)
lasso.fit(X_train, y_train)
lasso_pred = lasso.predict(X_test)
lasso.score(X_test, y_test)

Question 21

What is Ridge regression and how to use it

Answer 21

It is a minimize loss function where alpha is the hyperparameter that needs to be tuned. Low alpha is less complex and can lead to overfitting while very high alpha can lead to underfitting

ridge = Ridge(alpha=0.1, normalize=True)
ridge.fit(X_train, y_train)
ridge_pred = ridge.predict(X_test)
ridge.score(X_test, y_test)

Question 22

How to use Lasso for feature selection

Answer 22

lasso = Lasso(alpha=0.1)
lasso_coef = lasso.fit(X, y).coef_
print(lasso_coef)

Question 23

How to get the coefficient attribute

Answer 23

.coef_

Question 24

What does logistic regression do and how does it work

Answer 24

It outputs probabilities

If ‘p’ > 0.5, data labeled ‘1’
If ‘p’ < 0.5, data labeled ‘0’

Question 25

Accuracy formula

Answer 25

tp + tn / tp + tn + fp + fn

Question 26

Precision formula

Answer 26

tp / tp + fp

Question 27

Recall formula

Answer 27

tp / tp + fn

Question 28

What is a confusion matrix and how to use it

Answer 28

Matrix structured in the following way:

[[tp fn]
[fp tn]]

print(confusion_matrix(y_test, y_pred))

Question 29

What is a classification report and how to use it

Answer 29

Shows precision, recall, f1 score and support for the prediction set compared to the text set

print(classification_report(y_test, y_pred))

Question 30

Plotting ROC Curves

Answer 30

Idk

Question 31

What is GridSearchCV and how to use it and how to return results

Answer 31

Runs 2 parameters at a time across a grid to determine the best hyperparameters to use for the model

param_grid = {‘n_neighbors’ : no.arrange(1, 50)}
knn = KNeighborsClassifier()
knn_cv = GridSearchCV(knn, param_grid, cv=5)
knn_cv.fit(X, y)

knn_cv.best_params_
knn_cv.best_score_

Question 32

How to convert categorical features to numerical features for Scikit-learn and pandas

Answer 32

Use dummy variables for pandas: pd.get_dummies(df)

Use one-hot-encoder for Scikit-learn: OneHotEncoder()

Question 33

How to drop missing data

Dataframe: df
Columns to drop missing values in: insulin, triceps and bmi

Answer 33

df. insulin.replace(0, np.nan, inplace=True)
df. triceps.replace(0, np.nan, inplace=True)
df. bmi.replace(0, np.nan, inplace=True)

Each had 768 entries but now 394, 541 and 757 bc the rest wer dropped

Question 34

df.dropna()

Answer 34

Drops all rows containing 0 in the entire dataframe; too much data is lost