CIS 663 AIU Online Computer Science Questions

CIS 663 AIU Online Computer Science Questions

AIU Online Computer Science Questions

Description

CIS 663 Biometrics
Assignment #4
Due: Before Live session 8
For this assignment it is expected you will write some code. You may use
any language or even just an excel spreadsheet. But you must turn in your
code and provide instructions on how to run.
Question 1) Using the iris (plant not eye) data set found at the UCI ML
Database, using the just first four columns of data:
1. sepal length in cm
2. sepal width in cm
3. petal length in cm
4. petal width in cm
5. class:
— Iris Setosa
— Iris Versicolour
— Iris Virginica
Compute the 3 closest pairs of flowers using:
a) Euclidean distance
b) Manhattan distance

c) Cosine similarity
Recall that
 A · B = Ax × Bx + Ay × By + …
 ||A|| = √ Ax
2
+ Ay
2
+…
Question 2) Given the summarized following dataset use Naïve bayes to
compute which class the following inputs belong to.

Fruit Red Has Leaves Has Seeds # Examples
Strawberry 300 250 200 350
Apple 400 100 300 600
Pear 100 50 200 250
Total 800 400 700 1200
There are three types of fruit, strawberries, apples and pears. We’ve
analyzed 1200 total individual pieces of fruit and for each category recorded:
1. If it is red
2. If it had leaves
3. If it had seeds
And recorded the counts for each category. The # Examples tells us the
actual number of pieces of fruit of each type that were evaluated. For
example, of the 350 strawberries examined, 300 were red.
Classify the following inputs using Naïve Bayes, show the probability of each
class for full credit.
 Input 1: Not Red, No Leaves, Has Seeds
 Input 2: Red, No Leaves, No Seeds
 Input 2: Red, Has Leaves, Has Seeds
Question 3)
a) What you can say about this ROC curve? What does a point on the ROC
curve correspond to?
b) Compare two ROC curves. Which one is a better model and why?
c) Draw or describe what a ROC curve would look like for a random guess.
What would be varying over the set of random guess algorithms?
Explain why it looks the way it does.
Question 4) You are developing a finger authentication system for parking
garage. The system works as follows.
 User approaches the gate and punch in the ID
 Puts his/her finger on the screen
 Once a fingerprint matches one of the registered users, the
system charges the user $10.
 You know that about 50% of the requests come from an imposter
hoping to park their car for free.
 When a transaction is made, the system sends the user a receipt
via email.
 User has a quick option to say that this was not a valid
transaction, then the system needs to refund the money to the
user.
 You have found out that if you reject a legitimate user, they
immediately leave and find a different parking garage that they
will be using regularly, and your business loses $100 of potential
revenue.
which of the following algorithms you would like to use and describe
how you will set a threshold for the system. Why?
Question 5) You have a ECG data given in the table with all detected
fiducial points. Extract features RQ, RP’, RP, RL’, RS, RS’, RT, RT’.
Fiducial points t
L’ 8.6
P 8.9
P’ 9.1
Q 9.3
R 9.5
S 9.6
S’ 9.9
T 10.4
T’ 10.6