# Digit recognition project

The goal of this project is to develop a pattern recognition system to identify the digit being shown by a 5x7 dot matrix display. You may recall such displays are common in LED signboards and dot-‐matrix font was used for printing in some early printers. The 5x7 dot matrix patterns for digits 0-‐9 are shown below.

You will assume that the display is faulty and at random it switches a cell from black to yellow or vice a versa with probability of p (p<<1) thus generating some varia+ons in the display.

The tasks involved in this project are as follows:

- Generate 100 faulty displays for each digit using a uniform random number generator. The 1,000 5x7 patterns thus generated form your training data set. [Choose p = 0.10]
- Generate 20 more faulty displays for each digit. These 200 samples are your test data set.
- Representing each pattern as a 35-‐dimensional vector, train a naive Bayesian classiﬁer and determine its accuracy over the training and test data.
- Next, represent each pattern through a computed set of features. Some examples are:

- Number of on(yellow) cells along each row and column (Total of 12 features)
- Number of on cells in overlapping 2x2 windows (Total of 24 features)

5. Repeat Step Three using representations of Step 4.

Tabulate the performance using confusion matrices. Summarize your experience and other project details in the form of a final report.

You may want to repeat with two other values of p, 0.05 and 0.15, for Step 3. Tabulate the performance.

### Subjects List

**Database (48)**

**Accounting (39)**

**Economics (34)**

**Finance (54)**

**Programming (122)**

**Management (69)**

**Marketing (25)**

**Report Writing (52)**

**Engineering (79)**

**Statistics (39)**

**Business Studies(45)**

**Case Studies (15)**

**Networking (26)**

**UML (9)**

**Science (23)**

**Nursing (11)**

**MATLAB (17)**

**Psychology (6)**

**Math (20)**

**Data Structure (9)**

**Python (15)**

**HRM (25)**

**IT (21)**

**Computer Science (57)**

**Law (4)**

**Essay (3)**

**Data Analysis (2)**

**Big Data (3)**

**Philosophy (2)**

**History (1)**