Câu hỏi trắc nghiệm về học máy

What is the first step in the data preparation process?

• Data clustering

• Understanding and exploring data

• Applying machine learning model

• Evaluating accuracy

What is the main goal of data preparation?

• Increase the amount of data

• Convert data to image format

• Create a clean and properly structured dataset

• Reduce training time

How does data normalization help in machine learning?

• Increase data transmission speed

• Reduce typing errors

• Increase prediction model efficiency

• Protect user data

Which technique belongs to the data preparation stage?

• Regression analysis

• Splitting training and testing data

• Hyperparameter optimization

• Pruning decision trees

Why is it necessary to check data integrity before training a model?

• To increase the dataset size

• To ensure data is not missing or skewed

• To avoid memory waste

• To increase the number of attributes

What can happen if data is not prepared correctly?

• The machine learning model will be easier to understand

• Data will not need to be labeled

• Prediction results may be inaccurate

• The model does not need evaluation

What is the difference between normalization and standardization?

• There is no difference

• Normalization brings data to the range [0,1], standardization brings it to a normal distribution

• Normalization loses outlier data, standardization keeps it intact

• Normalization is only used for string data

If the data contains many missing values, which processing step should be taken?

• Increase the number of training loops

• Replace or remove missing values

• Use classification algorithms

• Convert data to text

In a dataset with the attribute age, with values ranging from 18 to 65. To normalize this attribute to [0,1], which method should be used?

• Z-score normalization

• Min-Max scaling

• One-hot encoding

• Log transformation

You discover an attribute with many unusual duplicate values. What should you do?

• Change the machine learning algorithm

• Ignore the data

• Analyze for data entry errors

• Increase the number of epochs