Question 1 A training set is a collection of data An individual datum point in this training set is called an instance, sample, or an observation True False Question 2 Machine learning algorithms are commonly employed to solve problems that A) are too complex for traditional approaches B) can not be precisely defined C) that have no known algorithm D) both A and C Question 3 Machine learning algorithms can be classified based on several criteria Which of the following is most dissimilar to the others Supervised learning Unsupervised learning Semisupervised learning Reinforcement learning Question 4 Machine learning algorithms can be classified based on several criteria What class of algorithms oppose batch learning algorithms Question 5 Logistic regression is an example of an algorithm that can only be used for regression tasks, and not classification tasks True False 6 Clustering algorithms, in general, work by first specifying the proper clusters that a small sample of instances belong to, then the algorithm follows suit for the remaining instances True False 7 A batch learning system must be trained on all the available data at once True False 8 The term offline learning describes which of the following scenarios full model trained, launch model, training continues in production partially train model, launch model, training continues in production full model trained, launch model, training ceases in production partially train model, launch model, training ceases in production Question 9 A better name for the term online learning is arguably out of core learning incremental learning non batch learning network learning Question 10 Which phrase best describes the notion of a model generalizing well A model that has been trained on a sample dataset that represents the true population well A model that performs well on unseen testing data, for which the true labels values are known A model that performs well on the training data, for which the true labels values are known A model that achieves high performance after training on a relatively small portion of the full dataset Question 11 Instance based learning algorithms have model parameters that must be trained True False Question 12 Which of the following is an example of an instance based learning algorithm Principal Component Analysis Neural Network K Nearest Neighbors Ordinary Least Squares Linear Regression Question 13 ( The optimal model parameters for ordinary least squares linear regression is found by the MSE cost function or by some other objective function Question 13 options maximizing, minimizing minimizing, maximizing maximizing, maximizing minimizing, minimizing Question 14 Training a model means to compute the optimal model parameters by finding the values that minimize a cost function that measures how bad the fit is Question 14 options True False Question 15 If your model is not making good predictions, then adding attributes might help Question 15 options True False Question 16 If your model is not making good predictions, then using a more powerful model might help Question 16 options True False Question 17 An insufficient quantity of training examples will cause the model to make poor predictions Question 17 options True False Question 18 It is crucial to use a training set that is representative of the cases you want to generalize to Question 18 options True False Question 19 If your training data is full of errors, outliers, and noise, then it is harder for the learning system to detect underlying patterns in the data It is often well worth the effort to spend time cleaning up your training data The truth is, most data scientists spend a significant part of their time doing just that Question 19 options True False Question 20 A learning system will only be capable of learning if the training data contains enough relevant features and not too many irrelevant ones Question 20 options True False 21 For complex models, if the training set is noisy or too small, then they are likely to detect patterns in the noise itself True False 22 A very popular performance measure called is the ratio of correctly classified entities to the number of all classified entities Question 23 The data set that the system learns from is called the data set 24 Applying machine learning techniques on large datasets in an effort to discover patterns that would otherwise be difficult to detect is called 25 A class of machine learning algorithms that compare unseen (i e new) data points to known data points is called learning 26 A class of machine learning algorithms that elucidate patterns in the training data and construct a predictive system based on these patterns is an example of learning 27 A machine learning algorithm that is trained using known labels, or responses, is called a learning algorithm 28 A machine learning algorithm that aims to predict the discrete group that an observation belongs to is an example of a problem 29 Suppose you are building a model to predict a persons age based on their hair color, weight, hair length, and number of scars on their body In this scenario, weight and hair color are examples of 30 Suppose you are building a model to predict a persons age based on their hair color, weight, hair length, and number of scars on their body Predicting a persons age (i e a numeric target) is an example task 31 An algorithm having data that is unlabeled is called an learning algorithm 32 An algorithm that detects groups of similar cats based on their whisker length, coat color, weight, length, and top speed is considered a algorithm 33 An algorithm with the goal of simplifying the data without losing vital information is classified as a algorithm 34 Detecting unusual credit card transactions is an example of an unsupervised learning task called detection 35 An algorithm is trained to detect never before seen types of ice cream from images This is an example of a detection algorithm 36 An algorithm that can deal with data that is only partially labeled is called a learning algorithm 37 Reinforcement learning involves an intelligent within an that senses and must and actions to take In response to taking these actions, or are received, which are used to determine the best or that will earn the most over time (Use the following word bank to complete the above sentenc penalties, strategy, select, stimuli, rewards, policy, agent, perform, reward, environment) 38 An algorithm that loads part of the data, runs a training step on that data, and repeats the process until all of the data has been trained on due to a complete dataset that cannot fit in a machines main memory at once is an example of learning 39 Some models must employ a measure of similarity in order to make a prediction such models fall under the category of learning algorithms 40 in learning, a model is built based on the training data, then the model is used to make predictions 41 If your training dataset is too small, then it may be unrepresentative of the true population as a result of chance this is called (two words) 42 Large samples can be unrepresentative of the true population if the sampling method is flawed This is called (two words) 43 When a model has the training data, then it has performed well on the training data but does not generalize well 44 is when a model is constrained in order to reduce the risk of overfitting 45 of a model on a training set occurs when the model is too simple to capture the underlying structure of the data 46 The generalization performance is measured by making predictions on new instances, for which the associated responses are known To do this, you should split your data into two sets before training any models these sets are called the set and the set 47 If the training error is low (i e , your model makes few mistakes on the training set) but the generalization error is high, it means that your model is the training data

The Answer is in the image, click to view ...

Answered step by step

Verified Expert Solution

Link Copied!

Question

1 Approved Answer

Posted on Jul 29, 2024

Question 1 A training set is a collection of data. An individual datum point in this training set is called an instance, sample, or an

Question 1 A training set is a collection of data. An individual datum point in this training set is called an instance, sample, or an observation.

	True
	False

Question 2

Machine learning algorithms are commonly employed to solve problems that:

	A) are too complex for traditional approaches.
	B) can not be precisely defined.
	C) that have no known algorithm.
	D) both A and C

Question 3

Machine learning algorithms can be classified based on several criteria. Which of the following is most dissimilar to the others?

	Supervised learning
	Unsupervised learning
	Semisupervised learning
	Reinforcement learning

Question 4

Machine learning algorithms can be classified based on several criteria. What class of algorithms oppose "batch learning" algorithms?

Question 5

Logistic regression is an example of an algorithm that can only be used for regression tasks, and not classification tasks.

	True
	False

6 . Clustering algorithms, in general, work by first specifying the proper clusters that a small sample of instances belong to, then the algorithm follows suit for the remaining instances.

	True
	False

7 . A batch learning system must be trained on all the available data at once

	True
	False

8 . The term "offline learning" describes which of the following scenarios:

	full model trained, launch model, training continues in production
	partially train model, launch model, training continues in production
	full model trained, launch model, training ceases in production
	partially train model, launch model, training ceases in production

Question 9

A better name for the term "online learning" is arguably:

	out-of-core learning
	incremental learning
	non-batch learning
	network learning

Question 10

Which phrase best describes the notion of a model "generalizing" well?

	A model that has been trained on a sample dataset that represents the true population well.
	A model that performs well on unseen testing data, for which the true labels/values are known.
	A model that performs well on the training data, for which the true labels/values are known.
	A model that achieves high performance after training on a relatively small portion of the full dataset.

Question 11

Instance-based learning algorithms have model parameters that must be trained.

	True
	False

Question 12

Which of the following is an example of an instance-based learning algorithm?

	Principal Component Analysis
	Neural Network
	K-Nearest Neighbors
	Ordinary Least Squares Linear Regression

Question 13 (

The optimal model parameters for ordinary least squares linear regression is found by _____ the MSE cost function or by _____ some other objective function.

Question 13 options:

	maximizing, minimizing
	minimizing, maximizing
	maximizing, maximizing
	minimizing, minimizing

Question 14

"Training a model" means to compute the optimal model parameters by finding the values that minimize a cost function that measures how bad the fit is.

Question 14 options:

	True
	False

Question 15

If your model is not making good predictions, then adding attributes might help.

Question 15 options:

	True
	False

Question 16

If your model is not making good predictions, then using a more powerful model might help.

Question 16 options:

	True
	False

Question 17

An insufficient quantity of training examples will cause the model to make poor predictions.

Question 17 options:

	True
	False

Question 18

It is crucial to use a training set that is representative of the cases you want to generalize to.

Question 18 options:

	True
	False

Question 19

If your training data is full of errors, outliers, and noise, then it is harder for the learning system to detect underlying patterns in the data. It is often well worth the effort to spend time cleaning up your training data. The truth is, most data scientists spend a significant part of their time doing just that.

Question 19 options:

	True
	False

Question 20

A learning system will only be capable of learning if the training data contains enough relevant features and not too many irrelevant ones.

Question 20 options:

	True
	False

21 For complex models, if the training set is noisy or too small, then they are likely to detect patterns in the noise itself.

	True
	False

22 A very popular performance measure called____is the ratio of correctly classified entities to the number of all classified entities.

Question 23 The data set that the system learns from is called the____data set.

24 Applying machine learning techniques on large datasets in an effort to discover patterns that would otherwise be difficult to detect is called

25 A class of machine learning algorithms that compare unseen (i.e. new) data points to known data points is called____learning.

26 A class of machine learning algorithms that elucidate patterns in the training data and construct a predictive system based on these patterns is an example of____learning.

27 A machine learning algorithm that is trained using known labels, or responses, is called a_____learning algorithm.

28 A machine learning algorithm that aims to predict the discrete group that an observation belongs to is an example of a____problem

29 Suppose you are building a model to predict a persons age based on their hair color, weight, hair length, and number of scars on their body. In this scenario, weight and hair color are examples of_____

30 Suppose you are building a model to predict a persons age based on their hair color, weight, hair length, and number of scars on their body. Predicting a persons age (i.e. a numeric target) is an example____task.

31 An algorithm having data that is unlabeled is called an ______ learning algorithm.

32 An algorithm that detects groups of similar cats based on their whisker length, coat color, weight, length, and top speed is considered a______algorithm.

33 An algorithm with the goal of simplifying the data without losing vital information is classified as a ________algorithm.

34 Detecting unusual credit card transactions is an example of an unsupervised learning task called _______detection.

35 An algorithm is trained to detect never-before-seen types of ice-cream from images. This is an example of a______detection algorithm.

36 An algorithm that can deal with data that is only partially labeled is called a _______learning algorithm.

37 Reinforcement learning involves an intelligent ______within an ______that senses ______and must_____and _____actions to take. In response to taking these actions,

______ or_____are received, which are used to determine the best _____or____that will earn the most____over time. (Use the following word bank to complete the above sentenc: penalties, strategy, select, stimuli, rewards, policy, agent, perform, reward, environment)

38 An algorithm that loads part of the data, runs a training step on that data, and repeats the process until all of the data has been trained on due to a complete dataset that cannot fit in a machines main memory at once is an example of___learning.

39 Some models must employ a measure of similarity in order to make a prediction; such models fall under the category of____learning algorithms.

40 in_____learning, a model is built based on the training data, then the model is used to make predictions.

41 If your training dataset is too small, then it may be unrepresentative of the true population as a result of chance - this is called ____(two words).

42 Large samples can be unrepresentative of the true population if the sampling method is flawed. This is called ___(two words).

43 When a model has_____the training data, then it has performed well on the training data but does not generalize well.

44 ____is when a model is constrained in order to reduce the risk of overfitting.

45 ___of a model on a training set occurs when the model is too simple to capture the underlying structure of the data.

46 The generalization performance is measured by making predictions on new instances, for which the associated responses are known. To do this, you should split your data into two sets before training any models - these sets are called the____set and the ___set.

47 If the training error is low (i.e., your model makes few mistakes on the training set) but the generalization error is high, it means that your model is ___the training data.