This is my implementation of Perceptron for exercise 1.4 of Learning from Data a Short Course
It is a Python implementation of the Perceptron algorithm.
- Create data set (it must be labeled)
- Calculate some function (f) which splits the data into two sections
- Generate some function that represents perceptron running which classifies the data
- Run the perceptron function
- Run the update function
$w(t+1)=w(t)+y(t)x(t)$ on some given incorrect point. If there are none we're done!- Increment the repitition counter
- Rerun 4 and 5 until complete
python perceptron.py --num-points 20
--num-points NUM_POINTS, -n NUM_POINTS
The number of points of data you would like to generate.
--start-weights START_WEIGHTS, -s START_WEIGHTS
A tuple containing the starting values for x and y you would like to use. Defaults to (1, 1)
--bias BIAS, -b BIAS The bias to use in the equation y=b+w1+w2. Keep in mind w could be negative. The bias must be
between -10 and 10
--pause-length PAUSE_LENGTH, -p PAUSE_LENGTH
How long to pause and show the graph between iterations of the perceptron.
--load-previous-data LOAD_PREVIOUS_DATA
Load a file with the contents of a previous f_data. Mainly used for debugging.
--draw-iterations If set to true it will draw every iteration of the algorithm. You can control how fast the
draw occurs with pause-length.
This is what it looks like when running in real time: https://youtu.be/A_jwin6IwYM
- The purple line
$h$ is not labeled
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent