This is the introductory part of the first chapter
1. Multiple choice question
1. For mobile operators to predict customer churn, which of the following machine learning methods can be used is more appropriate ( )
A. Single linear regression analysis
B. Association method
C. Clustering algorithm
D. Multi-layer feedforward network
Correct answer: A
2. Which of the following statements about machine learning is wrong? ( )
A. High-quality data, computing power and algorithms are essential for a machine learning project.
B. Deep learning is a class of advanced algorithms for machine learning that can process complex data such as images, sounds, and texts.
C. There are many machine learning algorithms, and the algorithms that appeared later have better performance than the algorithms that appeared earlier.
D. Machine learning can imitate human learning to a certain extent, and can enhance human decision-making ability.
Correct answer: C
3. Which of the following steps is not a preprocessing work required for machine learning ( )
A. Standardization of numerical attributes
B. Variable correlation analysis
C. Outlier analysis
D. Discuss analysis needs with users
Correct answer: D
4. Data preprocessing is very important to machine learning. The following statement is correct ( )
A. The effect of data preprocessing directly determines the quality of machine learning results
B. Data noise has no effect on the training of neural networks
C. For those who have The data of the problem can be deleted directly
D. Preprocessing does not need to spend a lot of time
Correct answer: A
5. Google News collects a lot of news every day, and uses the ( ) method to group these news into several categories of related news. Therefore, the same group of news events often belong to the same topic when searching, so they are displayed together.
A. Regression
B. Classification
C. Clustering
D. Association rules
Correct answer: C
6. The wrong statement during the implementation of machine learning projects is ( )
A. The selection of suitable algorithms for machine learning is generally difficult to determine before analysis
B. This is the work of data analysts and does not require users to participate
C. Many machine learning tools Follow certain machine learning steps
D. The process of machine learning is not simple
Correct Answer: B
7. Machine learning methods can traditionally be divided into ( ) categories.
A. 3
B. 7
C. 4
D. 2
Correct answer: A
8. Which of the following statements about machine learning engineers is correct ( )
A. No need to know certain relevant business knowledge
B. No need to be familiar with data extraction and preprocessing
C. Need certain data analysis actual project training
D. After training Competent in actual data analysis
Correct answer: C
9. Which development language is best for machine learning? ( )
A. HTML
B. Python
C. C
D. Java
Correct Answer: B
10. Mobile operators segment customers to design packages and marketing activities, which of the following machine learning methods can be used ( )
A. Bayesian classifier
B. Association method
C. Clustering algorithm
D. Multi-layer feedforward network
Correct answer: C
11. Which of the following statements about feature selection is correct?
A. The fewer features selected, the better
B. The selected features should reflect the differences between different things as much as possible
C. The more features selected, the better
D. The above statement neither right
Correct Answer: B
12. Establish a model to predict which type of data mining task a certain other variable value belongs to based on known multiple variable values ( )
A. Classification rules
B. Regression analysis
C. Clustering
D. Information retrieval
Correct Answer: B
13. Which is a reasonable definition of machine learning?
A. Machine learning is the science of computer programming
B. Machine learning learns from labeled data
C. Machine learning is the field that allows robots to act intelligently
D. Machine learning enables computers to Learn without explicit programming
Correct answer: D
14. The incorrect statement about machine learning preprocessing is ( )
A. Preprocessing is an indispensable part of machine learning
B. Preprocessing only needs to be done once
C. Preprocessing involves data cleaning, format conversion, and null values Filling and so on a lot of work
D. The process of preprocessing may itself be a machine learning process
Correct Answer: B
15. A computer program learns a task T from experience E and uses P to measure performance. Moreover, the performance P of T increases with the increase of experience E. Suppose we feed a learning algorithm a lot of historical weather data and let it learn to predict the weather. What is a reasonable choice for P?
A. The probability of correctly predicting the weather for a future date
B. The task of weather forecasting
C. The process of computing large amounts of historical weather data
D. None of the above
Correct answer: A
16. ( ) is a machine learning algorithm that can be used to label data.
A. Clustering algorithm
B. Association rule algorithm
C. None of the above
D. Regression algorithm
Correct answer: D
17. Victor Meyer-Schönberger continued to emphasize a point of view in the book "The Era of Big Data: A Revolution in Life, Work, and Thinking": the advent of the era of big data prevents us from artificially discovering data At the same time, we should pay more attention to the correlation relationship in the data rather than the causal relationship. Among them, the correlation between data can be directly mined by which of the following algorithms ( ).
A. K-means
B. Bayes Network
C. C4.5
D. Apriori
Correct answer: D
18. What is the difference between a regression problem and a classification problem?
A. Regression problems and classification problems require different input attribute values.
B. Regression problems have labels, but classification problems do not.
C. Regression problems have continuous output values, while classification problems have output values It is discrete
D. The output value of the regression problem is discrete, and the output value of the classification problem is continuous
Correct answer: C
19. Which machine learning models are trained to make a set of decisions based on the rewards and feedback they receive for their actions
A. All of the above
B. Supervised learning
C. Reinforcement learning
D. Unsupervised learning
Correct answer: C
20. The 2006 IEEE International Conference on Data Mining selected the 10 most important algorithms in the field of data mining, among which the algorithms that do not belong to these 10 categories are ( ).
A. Apriori algorithm
B. Support vector machine
C. Adaboost algorithm
D. Convolutional neural network
Correct answer: D
21. The term machine learning is defined by ( )?
A. None of the above
B. Arthur Samuel
C. James Gosling
D. Guido van Rossum
Correct Answer: B
22. Which of the following statements is correct ( ).
A. Classification and clustering are both supervised learning
B. Classification and clustering are both unsupervised learning
C. Classification is supervised learning and clustering is unsupervised learning D. Classification is unsupervised learning
and clustering is unsupervised learning Classes are guided learning
Correct answer: C
23. ( ) is the part of machine learning that works with neural networks.
A. deep learning
B. artificial intelligence
C. A and B
D. none of the above
Correct answer: A
24. The statement about incorrect data quality is ( )
A. Wrong data may produce harmful results for decision-making
B. Because of the large amount of data, poor data quality will not have much impact on machine learning
C. Data preprocessing The important purpose of D is to improve the quality of machine learning results
D. Dirty data extracted from business systems need preprocessing for modeling work
Correct Answer: B
25. The following statement is correct ( )
A. The purpose of machine learning is to discover useful information from data
B. The main task of machine learning is to discover potential laws from data, so as to better assist decision-making or realize machine automation Action
C. Machine learning is just the discovery of patterns in data generated by computer simulation methods
D. Machine learning is to use visualization methods to display multi-dimensional information in data
Correct Answer: B
2. Multiple choice questions
26. The method of machine learning consists of several elements such as ( ).
A. Loss function
B. Optimization algorithm
C. Model
D. Model evaluation index
Correct Answer: ABCD
27. Which of the following learning problems does not belong to supervised learning? ( )
A. Clustering
B. Regression
C. Classification
D. Dimensionality reduction
Correct Answer: AD
3. Judgment question
28. It is a regression problem to judge benign or malignant according to the volume of the tumor and the age of the patient.
A. Right
B. Wrong
Correct answer: wrong
29. In most machine learning projects, the three steps of data collection, data cleaning, and feature engineering take up most of the time, while data modeling takes up less of the total time.
A. Right
B. Wrong
Correct answer: yes
30. Knowing the information of your friends, such as contacts who often send emails, or your Weibo friends, WeChat Moments, we can use the clustering method to automatically group friends, so that each group People know each other well.
A. Right
B. Wrong
Correct answer: yes