Integrative Programming 5

Question 1

Linear Regression

Answer 1

a simple algorithm that models linear relationship between inputs and continous numerical output variable

Question 2

Linear Regression Use Cases

Answer 2

• stock price prediction
• house price prediction
• customer lifeline value prediction

Question 3

Linear Regression Advantages

Answer 3

• explainable method
• interpretable results by its coefficients
• faster to train other than machine learning models

Question 4

Linear Regression Disadvantages

Answer 4

• assumes linear relationship between input and output data
• sensitive to outliers
• can underfit small, high-dimensional data

Question 5

Logistic Regression

Answer 5

an algorith that models linear relationship between input and categorical outputs ( 1 or 0)

Question 6

Logistic Regresssion Use Cases

Answer 6

• credit risk score prediction
• customer churn prediction

Question 7

Logistic Regresssion Advantages

Answer 7

• interpretable and explainable
• less prone to overfitting when using regularization
• applicaple for multi-class prediction

Question 8

Logistic Regresssion Disadvantages

Answer 8

• assumes linear relationship between input and output
• Can overfit small, high-dimensional data

Question 9

Lasso Regression

Answer 9

Part of the regression family — it penalizes
features that have low predictive outcomes by
shrinking their coeOcients closer to zero. Can
be used for classification or regression

Question 10

Lasso Regression Use Cases

Answer 10

• predicting house prices
• predicting clinical outcomes based on health data

Question 11

Lasso Regression Advantages

Answer 11

• less prone to overfitting
• can handle high dimensional data
• no need for feature selection

Question 12

Lasso Regression Disadvantages

Answer 12

can lead to poor interopreability as it can keep highly correlated variables

Question 13

Ridge Regression

Answer 13

Part of the regression family — it penalizes
features that have low predictive outcomes by
shrinking their coeOcients closer to zero. Can
be used for classification or regression

Question 14

Ridge regression

Answer 14

• predictive maintenance for automobiles
• sales revenue prediction

Question 15

Ridge Regression Advantages

Answer 15

• less prone to overfitting
• best suited where data suffer from multicollinearity
• explainable and interpretable

Question 16

Ridge Regression Disadvantages

Answer 16

• all predictors are kept in the final model
• doesnt peform feature selection

Question 17

Ridge Regression Case Uses

Answer 17

• prediction for automobile maintenace
• sales revenue prediction

Question 18

Decision Tree

Answer 18

make decision rules on the features to produce predictions. It can be used for classification or regressio

Question 19

Decision Tree Use Cases

Answer 19

• customer churn prediction
• credit score modeling
• disease prediction

Question 20

Decision Tree Advantages

Answer 20

• explainable and interpretable
• can handle missing values

Question 21

Decision Tree Disadvantages

Answer 21

• prone to ovefitting
• sensitive to outliers

Question 22

Random Forest

Answer 22

learning method that combines the output of multiple decision trees

Question 23

Random Forest Use Cases

Answer 23

• credit score modeling
• house prices prediction

Question 24

Random Forest Advantages

Answer 24

• reduces overfitting
• higher accuracy than other models

Question 25

Random Forest Disadvantages

Answer 25

* not very interpretable * higher complexity training

Question 26

Gradient Boosting Regression

Answer 26

employs boosting to make predictive models from an ensemble of weak predictive learners

Question 27

Gradient Boosting Regression Used Cases

Answer 27

* prediction of car emission * prediction of ride hailing fair amount

Question 28

Gradient Boosting Regression Advantages

Answer 28

* better accuracy than other models * handle multicollinearity * can handle non linear relationships

Question 29

Gradient Boosting Regression Disadvantages

Answer 29

* sensitive to outliers therefore resulting to overfitting * computationally expensive and high complexity

Question 30

XGBoost

Answer 30

* effiecient and flexible * can be used for classification or regression

Question 31

XGBoost Used Cases

Answer 31

* claims in processing insurance * churn prediction

Question 32

XGBoost Advantage

Answer 32

* claims in processing insurance * churn prediction

Question 33

XGBoost Disadvantage

Answer 33

* hyperparameter tuning complexity * cannot perform well on spare datasets

Question 34

LightGBM Regressor

Answer 34

boosting framework designed to be more efficient other than implementation

Question 35

LightGBM Regressor Used CAses

Answer 35

* predicting flight time for airlines * predicting cholesterol level based on health data

Question 36

Tree-based model

Answer 36

1. decision tree 2. random forest 3. gradient boosting regression 4. XGBoost 5. LightGBM Regressor

Question 37

LightGBM Regressor Advantages

Answer 37

* can handle large amounts of data * computationally efficient * low memory usage

Question 38

LightGBM Regressor Disadvantages

Answer 38

* can overfit data due to high complexity * hyperparameter tuning can be complex

Question 39

Clustering Models

Answer 39

1. k-means 2. hierarical clustering 3. gaussian mixture model

Question 40

K-means

Answer 40

K-Means is the most widely used clustering approach—it determines K clusters based on euclidean distances

Question 41

K-Mean Used Cases

Answer 41

* customer segmentation * recommendation system

Question 42

K-Mean Advantages

Answer 42

* can handle scale large datasets * simple to interpret and implement * results to tight clusters

Question 43

K-Mean Disadvantages

Answer 43

* equires the expected number of clusters from the beginning * Has troubles with varying cluster sizes and densities

Question 44

hierarical clustering

Answer 44

* bottom up approcah where datasets are considered as their own clusters * merge of two clusters iteratively

Question 45

hierarical clustering used cases

Answer 45

* fraud detection * document clustering based on similarities

Question 46

hierarical clustering advantages

Answer 46

* no need to spcify number of clusters * resulting dendrogram is informative

Question 47

hierarical clustering disadvantages

Answer 47

* not always the best in clustering * cannot handle large datasets

Question 48

gaussian mixture modeling

Answer 48

probalistic modeling normally distributed clusters without dataset

Question 49

gaussian mixture modeling advantages

Answer 49

* compares probability for an observation belonging to a cluster * can identify overlapping clusters * provide better accurate results than k-means

Question 50

gaussian mixture modeling disadvantages

Answer 50

* requires complex tuning * requires setting the number of mixture components or clusters

Question 51

gaussian mixture modeling used cases

Answer 51

* customer segmentation * recommendataion system

Question 52

Apriori algorithm

Answer 52

Rule based approach that identifies the most frequent itemset in a given dataset where prior knowledge of frequent itemset properties is used

Question 53

Apriori algorithm used cases

Answer 53

* prodcut placement * recommendation engine * promotional optimization

Question 54

Apriori algorithm advantages

Answer 54

* results are intrepretable and intuitive * exhaustive approach as it finds all rules