answer for both questions: 4. Handwritten Digit Recognition. The goal is to recognize the digit in each
Fantastic news! We've Found the answer you've been seeking!
Question:
answer for both questions:
Transcribed Image Text:
4. Handwritten Digit Recognition. The goal is to recognize the digit in each image of the dataset given in "Digits Training" which contains some digits from the US Postal Service Zip Code. We are going to decompose the big task of separating ten digits into smaller tasks of separating two of the digits (binary classification). Use two digits: the final number in your UD ID and conveniently choose any other number to replicate the results from the slides chapter "The Learning Problem" (take into account the features that you are going to use for classification to choose the second number). (a) Extract 2 features from the images: average intensity and symmetry. Using this two features, implement the Perceptron Learning Algo- rithm. Use an error metric for binary classification. To compute Eout, use the testing set given to you in "DigitsTesting". Show only 200 iterations. (b) Repeat item (a), for the pocket algorithm. Show the same plots that are in Slide 32 and 33 of the chapter "The Learning Problem", that is, compare errors (Ein, Eout) and classification boundaries of the simple perceptron and the pocket algorithms. 4 images are expected. (c) Extract one more feature from the images that could help to improve your previous results. Describe how you compute this feature and why is it representative of your data?. (d) Repeat items (a) and (b), using the three features (average intensity, symmetry and the one that you choose in (c)). Hint: The classifica- tion boundary would be a plane given a 3D feature space. Dataset description: The first column in Digits Training and Digits Test- ing corresponds to the digit number, following columns correspond to 256 pixels of the 16 x 16 pixel image of the digit. Thus, we have 7291 inputs in Digits Training and 2007 inputs in Digits Testing. From these datasets, work only with those inputs that correspond to the digits you chose. Re- member, one of the digits corresponds to the final number in your UD ID. 4. Handwritten Digit Recognition. The goal is to recognize the digit in each image of the dataset given in "Digits Training" which contains some digits from the US Postal Service Zip Code. We are going to decompose the big task of separating ten digits into smaller tasks of separating two of the digits (binary classification). Use two digits: the final number in your UD ID and conveniently choose any other number to replicate the results from the slides chapter "The Learning Problem" (take into account the features that you are going to use for classification to choose the second number). (a) Extract 2 features from the images: average intensity and symmetry. Using this two features, implement the Perceptron Learning Algo- rithm. Use an error metric for binary classification. To compute Eout, use the testing set given to you in "DigitsTesting". Show only 200 iterations. (b) Repeat item (a), for the pocket algorithm. Show the same plots that are in Slide 32 and 33 of the chapter "The Learning Problem", that is, compare errors (Ein, Eout) and classification boundaries of the simple perceptron and the pocket algorithms. 4 images are expected. (c) Extract one more feature from the images that could help to improve your previous results. Describe how you compute this feature and why is it representative of your data?. (d) Repeat items (a) and (b), using the three features (average intensity, symmetry and the one that you choose in (c)). Hint: The classifica- tion boundary would be a plane given a 3D feature space. Dataset description: The first column in Digits Training and Digits Test- ing corresponds to the digit number, following columns correspond to 256 pixels of the 16 x 16 pixel image of the digit. Thus, we have 7291 inputs in Digits Training and 2007 inputs in Digits Testing. From these datasets, work only with those inputs that correspond to the digits you chose. Re- member, one of the digits corresponds to the final number in your UD ID.
Expert Answer:
Related Book For
International Marketing And Export Management
ISBN: 9781292016924
8th Edition
Authors: Gerald Albaum , Alexander Josiassen , Edwin Duerr
Posted Date:
Students also viewed these programming questions
-
PLEASE GIVE CORRECT ANSWERS Prove that the number of comparators in any sorting network is (n log n). [4 marks] (ii) What does Part (d)(i) imply in terms of the depth of any sorting network? [1 mark]...
-
"internet radios" for streaming audio, and personal video recorders and players. Describe design and evaluation processes that could be used by a start-up company to improve the usability of such...
-
PROJECT SUMMARY: You have been asked to submit a proposal to a client, Sara Johnson, who is moving the small firm to a new office location. The proposal is on the analysis and design of the office...
-
The following are hypothetical situations that require the choice of one of the two bases of measurement: cost or fair value. 1. You purchase your textbooks at the university bookstore to use during...
-
Does a business you work for or have worked for have one or multiple lines of business? List the line(s).
-
Abbott Industries is a well-known supplier of pharmaceuticals worldwide. Founded by Dr. Wallace Abbott, the company was incorporated in 1900 after he had been developing and making pharmaceuticals...
-
AnimPix, Inc., is a small company that creates computer-generated animations for films and television. Much of the company's work consists of short commercials for television, but the company also...
-
In the following exercises, use a contour consisting of a long rectangle with one side on the real axis and height 2 for Exercise 6, height for Exercise 7. 7. ex/n dx, n=2, 3,... x ex sinh x dx
-
Janice and Richard Morgan are married and have no dependents (MFJ). Janice works as anemployee from Worldwide Publishing and also as a freelance writer. She called her freelancewriting business...
-
The spring of a spring gun has force constant k = 430 N/m and negligible mass. The spring is compressed 6.30 cm and a ball with mass 0.0350 kg is placed in the horizontal barrel against the...
-
Discuss the budgetary issue causes the most strife in all areas of a health care organization?
-
A railroad hopper car filled with sand is rolling with an initial speed of 17.5 m/s on straight, horizontal tracks. You can ignore frictional forces on the railroad car. The total mass of the car...
-
Part I: Image filtering Considering an image denoising problem, let x be a noise-free m m image degraded by an additive white Gaussian noise w. The observed corrupted image y is given by: y = x +w....
-
Given Net Income of $50 800 , Total Assets $658 000, Retained earnings $224 400. What is the Internal Growth Rate?
-
The creation of S corporation as an alternative available for business organization created a variety of complications in the tax code and regs. One of these is the need for the accumulated...
-
Briefly describe the following types of group life insurance plans: a. Group term life insurance b. Group accidental death and dismemberment insurance (AD&D) c. Group universal life insurance d....
-
How would you respond to a person making the following statement: 'as a general rule, the export market manager has few sources of information available for use in market selection?
-
What do you see as the future for the Internet as a basic advertising medium and/or as a means for implementing sales promotion in foreign markets?
-
In 2009, after four decades of rapid growth, China surpassed the US to become the worlds largest car market in number of vehicles sold (Waldmeir, 2010a). China is currently the largest single market...
-
Let \(\left\{X_{n}ight\}_{n=1}^{\infty}\) be a sequence of random variables such that for each \(n \in \mathbb{N}\), \(X_{n}\) has a \(\operatorname{GAmma}\left(\alpha_{n}, \beta_{n}ight)\)...
-
Let \(\left\{X_{n}ight\}_{n=1}^{\infty}\) be a sequence of random variables where \(X_{n}\) is an ExPo\(\operatorname{NENTIAL}\left(\theta+n^{-1}ight)\) random variable for all \(n \in \mathbb{N}\)...
-
Let \(\left\{X_{n}ight\}_{n=1}^{\infty}\) be a sequence of random variables where for each \(n \in \mathbb{N}, X_{n}\) has an BERNoulli \(\left[\frac{1}{2}+(n+2)^{-1}ight]\) distribution, and let...
Study smarter with the SolutionInn App