NEW
Font size
WorksheetsBSCS 4-1 - Elective 4 - Midterm Examination - 2023
Total questions: 35
Worksheet time: 18mins
In the context of Machine Learning, what does the "Data" element refer to?
The process of training a model
The mathematical formula used for prediction
The evaluation metrics used to measure model performance
The input information used to train and test a model
What does the "Loss Function" element in Machine Learning represent?
The algorithm used to learn from the data
The mathematical function that measures the model's performance
The process of training the model
The method used to evaluate the model's predictions
Which element of Machine Learning focuses on updating the model's parameters based on the provided data?
Learning Algorithm
Model
Task
Evaluation
Which element of Machine Learning involves choosing appropriate features from the data?
Data
Task
Learning Algorithm
Model
Which of the following describes the mathematical function that establishes the connection between input data and output data?
Training
Algorithm
Loss Function
Model
This concept describes the process of using input data along with corresponding output data to establish a relationship through Machine Learning, which can then be utilized to predict output for new input?
Unsupervised Learning
Supervised Learning
Task
Model
One of the elements of machine learning that involves identifying the specific objective or goal that needs to be achieved using the available data. It requires asking questions like "What task can be accomplished using the given data?" or "What can be learned from the data?”
Unsupervised Learning
Supervised Learning
Task
Model
Which element of machine learning is responsible for evaluating the performance of a model by measuring its accuracy in predicting the expected outcome?
Training
Algorithm
Loss Function
Model
It is a set of multiple numeric features. We use it as an input to the machine learning model for training and prediction purposes.
Targets
Features
Categorical numeric sets
Feature Vector
Why is machine learning considered a critical and transformative technology that plays a vital role in various aspects of our modern society?
It enables the automation and optimization of complex tasks, leading to increased efficiency
It enhances decision-making and strategic planning through advanced analytics and predictive modeling
It empowers businesses to make data-driven decisions and optimize processes.
It revolutionizes industries and drives innovation by leveraging advanced algorithms and computational power.
In machine learning data analysis, why is quality data considered essential for the efficient operation and continuous improvement of the models?
Quality data enhances the model's ability to generalize and make accurate predictions
Quality data helps the machine learning model to learn intelligently
Quality data will make the process of algorithm smooth and no interruptions
Quality data will make the output more complex
This type of data is sorted based on defining characteristics such as gender, social class, ethnicity, hometown, industry, or other labels.
Structured data
Categorical data
Quantitative data
Continuous data
Which of the following statements best describes unsupervised learning?
A machine learning technique where the model learns from labeled training data to make predictions or classify new, unseen data.
A machine learning technique that involves training a model without any labeled data
A machine learning technique where the model learns from unstructured data to uncover patterns and relationships
A machine learning technique used exclusively for supervised tasks
Which of the following statements correctly describes an advantage of supervised learning algorithms?
Supervised learning algorithms can generalize well to unseen data and make accurate predictions
Supervised learning algorithms are not affected by the quality or quantity of labeled training data
Supervised learning algorithms can work with unlabeled data and discover hidden patterns
Supervised learning algorithms require less computational resources compared to unsupervised learning algorithms
Which of the following correctly describes the purpose of the validation set in the process of supervised learning?
Used to train the model and optimize the weights and biases
Used to measure the performance of the trained model on unseen data and tune the hyperparameters
Used to evaluate the model's performance during the testing phase
Used to pre-process the data and remove outliers before training the model
In a supervised learning process, what is the primary purpose of the training set?
To measure the model's performance on unseen data
To fine-tune the model's hyperparameters
To evaluate the model's performance on labeled data
To optimize the model's weights and biases
During the training phase of supervised learning, when the model is exposed to labeled data and iteratively adjusts its internal parameter, what is the primary objective that drives this process?
To iteratively optimize the model's parameters, such as weights and biases, by minimizing the loss function, thereby aligning the model's predictions as closely as possible to the true values, and enhancing its predictive accuracy
To identify intricate patterns, complex relationships, and dependencies within the data, allowing the model to capture the underlying structure and nuances
To accurately classify the data into distinct and meaningful categories, leveraging the labeled examples to learn the discriminative characteristics that define each category
To assess and validate the performance of the model on unseen or held-out data, ensuring its ability to generalize well beyond the training set and effectively handle real-world instances that it has not been exposed to before
In a situation where the size of the training set is limited while the complexity of the model used is high, what is the anticipated outcome or behavior that is likely to occur?
The model will have high bias but low variance
The model will have low bias but high variance
The model will have high bias but low variance
The model will overfit the training data
Eager Learners in the context of classification refers to the algorithms that build a classification model using the entire training dataset before making any predictions or classifications. These learners eagerly construct a general hypothesis based on the provided training data and then use this hypothesis to classify new, unseen instances. Which of the following is an example of eager learner.
Case-Based Reasoning
Naïve Bayes
K-Nearest Neighbors
Locally Weighted Learning
Which of the following assumptions is NOT required for Simple Linear Regression?
Linearity: The relationship between the variables is linear
Independence: The residuals are independent of each other
Multicollinearity: The predictor variable is not highly correlated with other
Homoscedasticity: The variance of the residuals is constant across all levels of the predictor variable
Which of the following scenarios is suitable for Simple Linear Regression?
Predicting the type of disease based on patient symptoms
Estimating the price of a car based on its features (e.g., mileage of the car (kmpl), max power, maximum power of engine, etc.)
Identifying customer segments based on their purchasing behavior
Classifying emails as spam or non-spam based on their content
In the K-nearest neighbors’ algorithm, which method is typically used to determine the optimal value of K, representing the number of neighbors to consider for classification?
It is determined based on the number of features in the dataset
Based on the average distance between data points
This will be determined through trial and error
Using cross-validation or other model selection techniques
In the K-nearest neighbors’ algorithm, what happens if the number of neighbors (K) is set to a very large value?
The model becomes more sensitive to noise and outliers
The model becomes more prone to overfitting
The decision boundary becomes more flexible and complex.
The computational complexity of the algorithm increases significantly
In Naive Bayes, how are the likelihoods of the features computed?
By assuming a Gaussian distribution for the features
By summing the feature values for each class
By estimating the probability distributions of the features
By applying a logarithmic transformation to the feature values
The Naive Bayes algorithm involves calculating the posterior probability of a class given the feature values. How is this probability computed?
By multiplying the likelihoods of the features
By applying Bayes' theorem
By computing the prior probability of the class
By normalizing the probabilities
Known as a powerful supervised machine learning algorithm used for classification and regression tasks, which aims to find an optimal hyperplane in a high-dimensional feature space that separates the data points of different classes with the largest margin
K-Nearest Neighbors
Support Vector Machine
Logistic Regression
Naïve Bayes
These are considered as data points that lie closest to the decision boundary. They play a crucial role in defining the hyperplane in support vector machine algorithm.
Margins
Feature Vectors
Feature Spaces
Support Vectors
In K-Nearest Neighbor (KNN), there are different distance metrics we can use. Which of the following distance metrics is used to determine the proximity between a new data point and the existing labeled data points? This also considered as the most used distance metric in KNN, especially for continuous numerical features.
Minkowski distance
Manhattan distance
Euclidean distance
Hamming distance
A term used in an algorithm to map the input data to a specific category.
Classifier
Target
Initialize
Cluster
During the training phase of machine learning, what is the typical ratio in which data is split for training and testing?
70/30
80/20
90/10
60/40
What type of data is fed into the model during the training phase?
Only input data
Only output data
Both input and output data
None of the above
What does it mean when we say the model "learns" during the training phase?
The model memorizes the training data
The model applies pre-defined rules to the data
The model performs calculations on the training data
The model builds its own logic based on the training data
What is the purpose of splitting the data into training and testing sets?
To evaluate the model's performance on unseen data
To prevent the model from overfitting the training data
To measure the accuracy of the model's predictions
All of the above
What is the purpose of splitting the data into training and testing sets?
To evaluate the model's performance on unseen data
To prevent the model from overfitting the training data
To measure the accuracy of the model's predictions
All of the above
Suppose you have a dataset of 100 samples, and you decide to split it into training and testing sets using an 80:20 ratio. How many samples will be included in the training set?
80
20
60
40
