-
Notifications
You must be signed in to change notification settings - Fork 0
/
project_instructions.txt
39 lines (26 loc) · 1.08 KB
/
project_instructions.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Instructions
Implement a 1-layer neural network recognizing the
language of a given text .
Neural network has as many neurons as the number of unique languages
in a dataset. Neural network should recognize this
number automatically based on the training data.
Input vectors:
Input vector represents the proportion of each ascii
letter in a given text.
Output:
Each neuron calculates its linear output (net).
Use the maximum selector to find which neuron is activated (1)
and assume that other neurons are not activated (0).
Training:
Delta rule can be used.After each training
step weight vectors may be normalized to improve classification.
Data:
Create a training dataset by yourself. Create 3-4 separate
folders and name them to represent
languages (pl, en, de ...) . Each language folder
contains text files (10+) in a specified
language - one file contains a few paragraphs of text.
Make sure that the language you choose can be represented by ascii.
Testing:
Provide an interface to input a short text
that will be classified by your program.