You are on page 1of 7

Question Bank on Al-XI|

Chapter-6
Capstone Project
Learning Objectives:

After going through this chapter learners will be able to know:


Understanding the Capstone Project
Understanding/Defining the Problem
Problem Decomposing using Design Thinking Framework
Using an Analytical Approach
Metrics of Model Quality-Loss Function
Model validation

Chapter Flow Diagram:

Achiever Books
Question Bank on Al-XI|

Meaning of Capstone project


Examples of Capstone project
Problem
definition/understanding
Capstone
project Problem decomposition through
Design Thinking Framework
Business
Understanding
Analytic Approach
Data Requirements
Data collection
Analytics approach
through "Data Data understanding
Science
Stages involved in Methodology Data preparation
capstone project Modelling
Evaluation

Deployment
Feedback

Model Train Test Split Evaluation


validation Cross-Validation
Procedure

Mean Squared error


Regresssion
Metrics of Model losses Root Mean Square
Quality-Loss Error
Function Hyperparamete.
Classification losses

Important Points and Definitions:


1. Understanding Capstone-It can defined as the summarized learning experience got from integrating all
the learning's derived, from the program at one place in the form of a project is called as capstone project.

2. About capstone project-Capstone project is made with the intention that the students were able to map
real life problems in order to find asolution of the problems gathering all their experiences.
3. Examnplesof capstone project in AI-Some of the capstone project ideas are:
Stock prices predictor
Develop a sentiment analyzer
Fraud detection using machine learning
Brain tumor detection
Students results predictor

Page 58
verBooks
Question Bank on Al-XI|

4. Steps involved in Al project: Among all the technologies available today artificial intelligence is the
most transformative teclhnology. Thus every Al project follows the following six steps.
Problem definition or understanding the problem
Data gathering
Feature definition
AI modelling
Evaluation
Deployment
5. Understanding the problem- To solve a problem with the help of AI, there must be
a pattern. Unless
there is a pattern the problem cannot be solved with AI technology. So instead of jumping to solve a
problem it is very much necessary to embark on an AI development journey.
6.Decomposing the problem through DT framework-To solve a problem it is very much essential
understand the needs of pupil and reformulating with a brain storming to
ideas by "Design Thinking".
7. Steps involved in problem
decomposition-Problem decomposition steps are follows:
A-Understand the problem and express it in your own words.
Figure out the desired input and outputs.
Ask different questions for clarification.
B-Break the problem into several big parts and keep a note on this on a paper.
C-Divide the complicated parts into smaller pieces and
keep on doing this until all parts are small.
D-Try to code one small piece at a time.
Analyze on its implementations
Write the code
Test the code
Fix problem (if any found)
8. Analytics approach: While working in AI
questions, through data models domain or machine learning, to solve
gain deep insight actions or to are built to predict outcome or discover problems and answer
improve future outcomes analytics underline patterns. In order to
9. Stages in approach is followed.
foundational methodology
business understanding that for data science: Every
project irrespective of its size
lays the foundation for successful resolution of the starts with
Various stages involved are
business problems.

Achiever Books
Question Bank on Al-XI|

Business Analytic
understanding approach

Feedback Data
requirements

Deployment Data
collection

Evaluation Data
understanding

Modeling Data
preparation
10. Model validation: There are two
types of validation methods applied to a
model to check itefficiency.
Train test split evaluation (The train test split
for classification and technique can be used to test machine algorithms
regression problems.
Cross validation (Cross-validation is a re
models on a small sample of data. sampling technique for evaluating machine learning
11. Hyper parameters: These are
the parameters that are
before applying a explicitly defined to control the learning process
complexity machine-learning
of the model. Some of the algorithm to a dataset. Specifically, used to learn the capacity and
as Batch size, learning rate, hyper parameters are used for the
models, suchoptimization
etc. etc., and some are specific to the of the models, such
as Number of Hidden
layers,

Solved Question Bank


Section-A
Objective Type Questions
I. MCQS
Q-1 Every Al project must
undergo ---steps?
a. 3
b. 4 C. 5
Ans. d d. 6

0-2 Which of the


following statement is true about design
a. It provides a thinking methodology
solution-based approach to solve problems
b. It is useful to tackle complex
problems
c. It is a practical approach to protyping
d. All of the above

Achiever Books
Page 60
Question Bank on Al-XII

Ans, d

0-3 How many stages are there in design thinking framework.


b. 4 C. 5 d. 6
a. 3

Ans. C

C-4A researcher wants to study the association between gender and using a mobile phone. Data collected
for this study willbe [CBSE SQP]

a. Qualitative data b. Quantitative data c. Continuous data d. Classified data

Ans.- b

Q-5 Primary way to collect DATA (Data Gathering process)? [CBSE SQP]

a. Experiment b. Survey c. Interview d. Observation

Ans. B

Q-6 Which of the following is not a regression loss function [CBSE SQP]
b. Mean absolute error c. Mean square error d. Quantile loss
a. Exponential loss
Ans. a

0-7 The data scientist will use for predictive modelling? [CBSE SQP]

a. Artificial Intelligence b. Machine Learning

c. Training Set d. Deep Learning

Ans. C

Q-8 Which one does NOT belong with Classification loss? [CBSE SQP]

a. Log loss b. Mean Absolute Error c. Exponential Loss d. Hinge Loss

Ans. b

Q-9Which process does NOT come under Capstone Project? [CBSE SQP]
c. Deployment d. Data Gathering
a. AI Model b. AI Project Cycle
Ans. b

Q-10 Which one does NOT belong with Regression loss? [CBSE SQP]

a. Log Loss b. Mean Absolute Error c. Log cosh Loss d. Quantile Loss

Ans. a
Q11Adding a non-important feature to a linear regression model may result in. [CBSE SQP]

1)Increase in R-square
[CBSE SQP]
2)Decrease in R-square
a. Only 1 is b. Only2 is correct c. Either 1or 2 d. Neither 1 nor 2
correct
Achiever Books Page 61
Question Bank on Al-XI|

Ans. a

Q-12 Which of the following options is/are true for K-fold cross-validation? [CBSE SQP]
1) Increase in K willresult in higher time required to cross validate the result.
2) Higher values of Kwill result in higher confidence on the cross-validation result ascompared to lower
value of K.

3) If K=N, then it is called Leave one out cross validation, where Nis the number of observations
a. 1 and2 b.2 and 3C. 1and 3 d. 1, 2 and 3
Ans. d

Q-13 Choose Correct Option


a. DataRequirements >> Data Collection >> Data understanding >> Data
preparation
b. Data Requirements >> Data understanding >> Data
Collection>> Data preparation
c. Data Requirement >> Data Deployment>> Data Collection >> Data Gathering
d. Data Collection >> Data Request >> Data
Filtering >> Data Evaluation
Ans. a

Q-14 In AI development which framework is used?


a. Scikit-learn b. Tensorflow
c. PyCharm d. Matplotlib
Ans. b

Q-15 If your Neural Network model


seems to have high variance, what of the
promising things to try? following would be
a. Make the Neural Network
deeper b. Get more training data
c. Get more test data
d. Increase the number of units in each hidden
Ans. b layer

Q-16 Which stage in Design Thinking


missing [Prototype, Ideate, Test, Define] [CBSE
a. Evaluation
b. Empathies
Sample Paper]
c. Evolution d. Enrichment
Ans. b

Q-17 Which language is Most suitable for


developing AI?
a. Kotlin b. Swift
c. Python d. HTML
Ans.

0-18 A random sample of n=6


taken irom the population has the elements 6, 10, 13,
which option is False? l4, 18, 20. Then,
a. Point estimate for population mean is 13.5

Achiever Books
Page 62
Question Bank on Al-XII

estimate for population standard deviation is 4.68


b. Point

c. Point
estimate for population standard deviation is 3.5
mean is 1.01
d Point estimate for standard error of

Ans. C

O-10 The following data is used to apply a linear regression algorithm with least squares regression line
V=a1X. Then, the approximate value of ai is given by:(XIndependent variable, Y-Dependent variable)
1 20 30 40
400 800 1300

a. 27.876 b. 32.65o C. 40.541 d. 28.956

Ans. d
The volume of test data can be large,
Q-20 Which of the following statements is/are INCORRECT i)
the AI and ML algorithms keeping model
which presents complexities. ii) Your testing team should test mind. ii) Test data should include all
validation, successful learn ability, and algorithm effectiveness in
system.
irrelevant subsets of training data, i.e., the data youwill use for training the AI
a. None of the Below b. ii) c. ii) d. i),ii) & iii)

Ans. C

Q-21Which of these is NOT analytic based on type of question?


d. Data evaluation
a. Descriptive b. Statistical Analysis c. Forecasting

Ans. d
gender and using a mobile phone. Data
Q-22A researcher wants to study the association between
collected for this study will be----. [CBSE Sample Paper]
b. Quantitative data c. Continuous data d. Classified data
a. Qualitative data
Ans. b

Q-23If a models RMSE value is 160 then


a. It is a bad model
b. The model performance is high

c. There many mistakes in the model d. The model is average

Ans. b
package is used to
Q-24 Which of the following function in the scikit-learn python machine learning
Implement train-test split evaluation procedure?
a. Train_test() b. Test_split() c. Train_split() d. Train_ test_split)
Ans. D

-25 HoW many types of validation methods are there?

Achiever BookS Page 63

You might also like