WorksheetsML-MIDTERM-20251
Total questions: 60
Worksheet time: 10mins
What is artificial intelligence primarily defined as?
The development of computer hardware
The simulation of human intelligence by machines
The creation of robots only
The process of data collection
What is Predictive Analytics primarily used for?
Describing historical data
Forecasting future outcomes using data
Cleaning data for analysis
Encrypting sensitive data
Which of the following is considered a feature in a dataset?
Purchased: Yes/No
Age, income, credit score
Model prediction
Evaluation metric
What does the label in a dataset represent?
Independent variable
Algorithm used
Dependent variable or output
Training process
Which of these is NOT a core component of AI?
Computer vision
Natural language processing
Database management
Speech recognition
What is machine learning?
A type of computer programming
A method of hardware optimization
c. A branch of AI focusing on data and algorithms to imitate human learning
d. The process of building physical robots
A branch of AI focusing on data and algorithms to imitate human learning
The process of building physical robots
In supervised learning, what is essential?
Unlabeled datasets
Random data collection
No human intervention
Labeled datasets
What is the primary function of collaborative filtering?
To block unwanted content
To organize data alphabetically
To filter items based on similar users' reactions
To speed up computer processing
What is a key characteristic of unsupervised learning?
It requires labeled datasets
It needs constant human supervision
It discovers hidden patterns without human intervention
It only works with numerical data
What is Naive Bayes' key assumption?
All predictors are dependent
All predictors are independent
Only numerical data can be used
It requires human supervision
What percentage of data is typically used for training in the machine learning model discussed?
90%
70%
60%
80%
What is overfitting in machine learning?
When the model is too simple
When the model covers more than required data points
When the model has no variance
When the model is undertrained
What is a confusion matrix used for?
To confuse the machine learning model
To organize raw data
To define classification algorithm performance
To create neural networks
What does a True Positive (TP) in a confusion matrix indicate?
An incorrect negative prediction
An incorrect positive prediction
A correct positive prediction
A missed prediction
What is precision in machine learning metrics?
The speed of processing
The ratio of correct positive predictions to total positive predictions
The total accuracy score
d. The model's processing power
The model's processing power
What is predictive analytics?
Looking at past data only
Organizing current data
Using data to forecast future outcomes
Analyzing hardware performance
Which is NOT a way to avoid overfitting?
Cross-validation
Reducing training data
Early stopping
Regularization
What is underfitting characterized by?
High variance and low bias
Low variance and low bias
High bias and low variance
Perfect predictions
What is the command for showing the fist five record using pandas?
head()
tail()
select ()
range(5)
Which metric measures the proportion of actual positives correctly identified?
Recall
Specificity
c. Accuracy
Precision
What is content-based filtering primarily based on?
User demographics
Historical data only
Item features and user preferences
Random selection
What tool provides machine learning algorithms for classification and regression?
What is the primary purpose of cross-validation?
To increase processing speed
To ensure model reliability
To reduce data size
To create new data
Which type of learning does not require labeled data?
Supervised learning
Semi-supervised learning
Unsupervised learning
Reinforced learning
What is the primary challenge with text data from the internet?
It's mostly unlabeled
It's too structured
It's not accessible
It's too small
Which of the following is a supervised learning task?
Image compression
Customer segmentation
Spam email classification
Topic modeling
Which of these is an example of regression?
Fraud detection
Loan approval
Stock price prediction
Face recognition
What type of data does unsupervised learning use?
Labeled data
Unlabeled data
Both labeled and unlabeled data
None
Which algorithm is commonly used for clustering?
Decision Tree
Random Forest
K-Means
Naïve Bayes
Which type of learning involves reward and punishment to improve performance?
Supervised Learning
Unsupervised Learning
Reinforcement Learning
Deep Learning
What is the main goal of data preprocessing?
Collecting more data
Cleaning and transforming data
Building neural networks
Creating test data
In supervised learning, labels refer to:
Input variables
Independent variables
Target variables
Features
Which of the following is structured data?
Tweets
Images
Tables in a database
Audio recordings
Which of the following is unstructured data?
Gender column in Excel
Student grades table
Product reviews
Salary records
What is the purpose of training data?
To test model accuracy
To evaluate model performance
To help the model learn patterns
To clean data
Testing data is used to:
Train the model
Evaluate model generalization
Store old datasets
Build new models
Which of the following can cause underfitting?
Too complex model
Too simple model
Large dataset
Deep neural network
Which of the following is a symptom of overfitting?
High bias, low variance
High training accuracy, low test accuracy
Low training accuracy
Small dataset with poor results
Bias occurs when a model is ______.
Too simple
Too complex
Too accurate
Too large
Variance occurs when a model is ______.
Too simple
Too complex
Too accurate
Too large
The bias-variance tradeoff seeks to balance:
Data quality and data size
Simplicity and complexity
Accuracy and speed
Precision and recall
In classification, which metric measures the proportion of correctly predicted positive observations?
Precision
Recall
F1-score
Accuracy
Which step is NOT part of using Google’s Teachable Machine?
Collect and label images
Train the model
Write code manually
Test the model
Machine Learning differs from traditional programming because it:
Uses rules explicitly written by humans
Learns automatically from data
Doesn’t require data
Focuses only on robotics
Unstructured data includes which of the following?
Tables
Database records
Text, image, audio, and video data
Numerical data
What is the purpose of data preprocessing?
To create training models
To clean and organize raw data for analysis
To visualize the dataset
To store large amounts of data
What does data cleaning involve?
Formatting reports
Fixing or removing incorrect or duplicate data
Creating dashboards
Gathering new data
What is the training dataset used for?
Testing model accuracy
Visualizing patterns
Teaching the model to learn from data
Evaluating performance
What is the purpose of the testing dataset?
To build the model
To train the model
To evaluate model performance on unseen data
To clean the data
Underfitting happens when a model:
Memorizes the training data
Fails to capture underlying patterns in the data
Overlearns the noise in data
Uses too many features
Bias in machine learning refers to:
The model learning too much from noise
Error from overly simple assumptions
Large variability in predictions
Overtraining the model
Variance in machine learning refers to:
Model’s simplicity
Learning too little
Model learning too much from data noise
Missing values in data
Naive Bayes classifier assumes:
All features are dependent
Features are independent of each other
Data must be normalized
The model uses clustering methods
What is the function of model.predict() in Scikit-learn?
Train a model
Split data into sets
Predict labels for new data
Calculate accuracy
What does True Positive (TP) represent?
Predicted negative, actual positive
Predicted positive, actual positive
Predicted positive, actual negative
Predicted negative, actual negative
Accuracy is calculated using which formula?
(TP + FP) / Total
(TP + TN) / Total
(TP + FN) / Total
(FP + TN) / Total
Which performance metric measures how many of the predicted positives are actually positive?
Recall
Precision
F1-Score
Accuracy
The F1-Score combines which two metrics?
Accuracy and Recall
Precision and Recall
True Positive and False Positive
Sensitivity and Specificity
What is the formula for Precision?
TP ÷ (TP + FP)
TP ÷ (TP + FN)
TN ÷ (TN + FP)
FN ÷ (FP + TN)
What is the formula for Recall?
TP ÷ (TP + FP)
TP ÷ (TP + FN)
TN ÷ (TN + FP)
FP ÷ (FP + TN)
