Modelling Phase | CBSE Class 10 AI Project Cycle Notes – 5

1. What do you mean by modelling phase?

It is the stage in which AI models are created after mathematically analyzing the data and finding relationships among data.

2. What activities are performed during modelling phase?

i. AI models from data are created and best model among them is chosen.
ii. After choosing best model, AI algorithms are developed.
iii. Relationship among data is also defined.

3. What is the purpose or significance of modelling in Ai cycle?

It is used to identify the problem and training the AI model with the prepared data. It is also used for risk analysis, decision support and optimization of the AI model.

4. What do you mean by a rule based model?

A rule based model is a type of AI models that can make decisions and solve problems by using predefined rules and patterns.

5. Why are rule based models said to be static?

Rule based models are called static because they do not change with time.

6. What do you mean by learning based model?

Learning based model is a type of AI model that can learn from a large amount of data. There are not predefined rules and patterns defined by a developer. Machine learns with each new data and generates own algorithms.

7. Why are learning based system called dynamic?

Learning based models are called dynamic because they continuously learn and improve their performance with each new input.

8. Define Machine Learning.

It is a branch of AI that enables machines to automatically learn and improve tasks with experience and by the use of data.

9. Define deep learning.

Deep learning is a subset of machine learning in which machine learns by filtering and cleaning the input data using by rule based algorithm. It can take complex decision by using layers.

10. What are the types of learning based models?

i. Supervised Machine learning
ii. Semi supervised machine learning
iii. Unsupervised machine learning
iv. Reinforcement learning

11. What do you understand by labelled and unlabeled data?

i. Data without any explanatory tags or names is called unlabeled data.
ii. Data with explanatory tags and names is called labelled data.

12.What is supervised Learning?

Supervised learning is a machine learning approach in which a machine learns from a labelled data along with and desired output.

13. What are supervised learning models?

OR

How supervised Learning model works?

In supervised learning, AI model is created by using labeled training data that consists of input data and expected output. Based on the training data, the model learns to produce the output. Feedback of every output is provided that helps the AI model to make changes to its algorithms if output doesn’t match with expected result. 70% of data is the Training data and 30% is testing data.

14. Which type of models have a set of expected outputs along with the input data?

Ans. Supervised learning

15. Write advantages of supervised learning.

i. It is less complex.
ii. It is highly accurate method.
iii. It is useful when user has experience of similar problems.

16. Write disadvantages of supervised learning.

i. It works with labelled data so data must be preprocessed.
ii. Computational time is very high.
iii. Training data must be based on real working examples.

17. Name two supervised learning techniques?

i. Regression
ii. Classification.

18.What do you mean by regression?

This is a supervised learning method that can find relationship between different variables produce an output.

19.What do you mean by classification?

This is a supervised learning method that can classify new unlabeled data on the basis of training data.

20. Give some examples of applications where classification supervised learning technique is useful.

Document classification, Malware classification, fraud detection and detecting spam mails.

21. Give some examples of applications where Regression supervised learning technique is useful.

Financial forecasting( like house price estimates or stock prices), sales and weather forecasting.

22. What do you mean by unsupervised learning model?

It is a type of machine learning that learns from data without human supervision.

23 How does an unsupervised learning models work?

Unsupervised learning models uses machine learning algorithms to analyze, find patterns among unlabeled data and group data without human intervention.

24 Which types of models do not get expected outputs along with the input data?

Unsupervised learning

25. Write advantages of unsupervised learning.

i. It is useful to find unknown patterns in data.
ii. It is highly useful in real work problems because it can work with unlabeled data.
iii. Training takes place in real time.

26. Write disadvantages of unsupervised learning.

i. Accuracy may be compromised.
ii. It may not produce expected outputs because it works with unlabeled data.

27. Write names of Unsupervised learning techniques.

i. Clustering
ii. Association
iii. Dimensionality Reduction

28. What do mean by clustering? Gives examples.

It is an unsupervised learning technique that groups unlabeled data on the basis of their similarities or differences.
Examples: It can be used for identifying fake news, grouping similar documents on the basis of various characteristics.

29. What do mean by Association ? Gives examples.

Association is unsupervised learning technique that finds important relations between variables or features in a data set.
Example: Association can be used to recommend products to customers on online shopping sites according to purchasing habits of customers.

30. What do mean by Dimensionally reduction? Gives examples.

Dimensionally reduction is the technique to reduce the number of input variables in training data but the meanings and sense of data should remain same. It is used for reducing complexity of a problem.
Example: It can be used in data visualization, video compression and human gene expression.

31. What do you mean by semi Supervised learning?

Semi Supervised learning is a combination of supervised and unsupervised learning techniques. It uses both labelled and unlabeled data for training for AI model.

32. What is reinforcement learning?

It is the technique in which machines learns using a system of reward and penalty. The learning system called agent learns in an interactive environment where agent performs different actions. It receives rewards for produding correct output and penalty for wrong output.

33 Explain the components of reinforcement learning model?

i. Agent: The learning system or application.
ii. Environment: Different components with which agent interacts.
iii. Policy : Rules followed by agent to take actions.
iv. Action: Activities performed by the agent.
v. Reward: It is signal received by the agent on taking different actions.

34. Explain different types of reinforcement learning models?

There are two types of reinforcement learning models:
i. Positive Reinforcement: It focusses on increasing rewards to encourage a particular behaviour.
ii. Negative Reinforcement: It focusses on removing penalty to encourage a particular behaviour.

35. Give some examples of applications of reinforcement learning models.

i. Training of self-driving cars.
ii. Training robots to perform dangerous tasks.
iii. Experienced based recommendation system.

36. Write advantages of reinforcement learning.

i. It focusses on the problem as a whole.
ii. It can work easily in unknown environment.
iii. It is useful to solve very complex problems.
iv. It can correct errors during the training.
v. It controls error repetition.

37. Write disadvantages of reinforcement learning.

i. It is not suitable for simple problems.
ii. It needs a big amount of data.
iii. it may take a long time.
iv. It is very expensive.