-
Notifications
You must be signed in to change notification settings - Fork 1
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
Showing
1 changed file
with
139 additions
and
0 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,139 @@ | ||
| #+title: Read Me Org File | ||
|
|
||
| This package was designed to present classification boundary methods. The three methods include broad initial sampling, low point-density sampling, and path tracing. The following code demonstrates how to use each module to run classification boundary simulations. Here we use a classification polygon to determine the classification boundary. | ||
|
|
||
| * Import Package from Github | ||
| This will import the boundary search package and all the sub-modules within it. It will also import matplotlib's Polygon and Path for use in the classification tasks. | ||
| #+BEGIN_SRC jupyter-python | ||
| import os | ||
|
|
||
| os.system("pip install git+https://github.com/mayabhat/boundary_search.git --upgrade") | ||
|
|
||
| from sbs.densitysampling import sample as dsam | ||
| from sbs.pathtracing import sample as psam | ||
| from sbs.broadsampling import sample as bsam | ||
|
|
||
| from shapely.geometry import Polygon | ||
| import matplotlib.path as mplPath | ||
| import matplotlib.pyplot as plt | ||
| import numpy as np | ||
| from sklearn import svm | ||
| #+END_SRC | ||
|
|
||
| #+RESULTS: | ||
| :RESULTS: | ||
| Collecting git+https://github.com/mayabhat/boundary_search.git | ||
| Cloning https://github.com/mayabhat/boundary_search.git to /private/var/folders/bd/rbb6rqr92vl2rjbn7_9b7lw80000gn/T/pip-req-build-447_7gfs | ||
| Running command git clone --filter=blob:none --quiet https://github.com/mayabhat/boundary_search.git /private/var/folders/bd/rbb6rqr92vl2rjbn7_9b7lw80000gn/T/pip-req-build-447_7gfs | ||
| Resolved https://github.com/mayabhat/boundary_search.git to commit e4e94b58816aa5ca3c70804bcb3f2545596d7fa8 | ||
| Preparing metadata (setup.py): started | ||
| Preparing metadata (setup.py): finished with status 'done' | ||
| Requirement already satisfied: autograd in /Users/mayabhat/opt/anaconda3/lib/python3.9/site-packages (from sbs==0.0.1) (1.4) | ||
| Requirement already satisfied: pyDOE2 in /Users/mayabhat/opt/anaconda3/lib/python3.9/site-packages (from sbs==0.0.1) (1.3.0) | ||
| Requirement already satisfied: sklearn in /Users/mayabhat/opt/anaconda3/lib/python3.9/site-packages (from sbs==0.0.1) (0.0) | ||
| Requirement already satisfied: matplotlib in /Users/mayabhat/opt/anaconda3/lib/python3.9/site-packages (from sbs==0.0.1) (3.4.3) | ||
| Requirement already satisfied: numpy in /Users/mayabhat/opt/anaconda3/lib/python3.9/site-packages (from sbs==0.0.1) (1.22.4) | ||
| Requirement already satisfied: pycse in /Users/mayabhat/opt/anaconda3/lib/python3.9/site-packages (from sbs==0.0.1) (2.1.5) | ||
| Requirement already satisfied: pandas in /Users/mayabhat/opt/anaconda3/lib/python3.9/site-packages (from sbs==0.0.1) (1.3.4) | ||
| Requirement already satisfied: shapely in /Users/mayabhat/opt/anaconda3/lib/python3.9/site-packages (from sbs==0.0.1) (2.0.1) | ||
| Requirement already satisfied: future>=0.15.2 in /Users/mayabhat/opt/anaconda3/lib/python3.9/site-packages (from autograd->sbs==0.0.1) (0.18.2) | ||
| Requirement already satisfied: cycler>=0.10 in /Users/mayabhat/opt/anaconda3/lib/python3.9/site-packages (from matplotlib->sbs==0.0.1) (0.10.0) | ||
| Requirement already satisfied: kiwisolver>=1.0.1 in /Users/mayabhat/opt/anaconda3/lib/python3.9/site-packages (from matplotlib->sbs==0.0.1) (1.3.1) | ||
| Requirement already satisfied: pillow>=6.2.0 in /Users/mayabhat/opt/anaconda3/lib/python3.9/site-packages (from matplotlib->sbs==0.0.1) (8.4.0) | ||
| Requirement already satisfied: pyparsing>=2.2.1 in /Users/mayabhat/opt/anaconda3/lib/python3.9/site-packages (from matplotlib->sbs==0.0.1) (3.0.4) | ||
| Requirement already satisfied: python-dateutil>=2.7 in /Users/mayabhat/opt/anaconda3/lib/python3.9/site-packages (from matplotlib->sbs==0.0.1) (2.8.2) | ||
| Requirement already satisfied: pytz>=2017.3 in /Users/mayabhat/opt/anaconda3/lib/python3.9/site-packages (from pandas->sbs==0.0.1) (2021.3) | ||
| Requirement already satisfied: scipy in /Users/mayabhat/opt/anaconda3/lib/python3.9/site-packages (from pycse->sbs==0.0.1) (1.7.1) | ||
| Requirement already satisfied: scikit-learn in /Users/mayabhat/opt/anaconda3/lib/python3.9/site-packages (from sklearn->sbs==0.0.1) (1.2.0) | ||
| Requirement already satisfied: six in /Users/mayabhat/opt/anaconda3/lib/python3.9/site-packages (from cycler>=0.10->matplotlib->sbs==0.0.1) (1.16.0) | ||
| Requirement already satisfied: joblib>=1.1.1 in /Users/mayabhat/opt/anaconda3/lib/python3.9/site-packages (from scikit-learn->sklearn->sbs==0.0.1) (1.2.0) | ||
| Requirement already satisfied: threadpoolctl>=2.0.0 in /Users/mayabhat/opt/anaconda3/lib/python3.9/site-packages (from scikit-learn->sklearn->sbs==0.0.1) (2.2.0) | ||
| :END: | ||
|
|
||
| * Define a Classification Polygon | ||
| #+BEGIN_SRC jupyter-python | ||
| polygon = Polygon([(1, 1),(1, 2),(2, 2),(2, 1)]) | ||
| x,y = polygon.exterior.xy | ||
|
|
||
| fig, ax = plt.subplots(figsize=(5,5)) | ||
| ax.plot(x, y) | ||
| ax.set_xlim((0, 3)) | ||
| ax.set_ylim((0, 3)) | ||
| ax.set_ylabel('x2') | ||
| ax.set_xlabel('x1') | ||
| plt.show() | ||
| #+END_SRC | ||
|
|
||
| #+RESULTS: | ||
| :RESULTS: | ||
| [[file:./.ob-jupyter/fa6d347b7db28f12a3789cd01e7223468f16d47d.png]] | ||
| :END: | ||
| Within this polygon will be class 1, and outside will be class 2. The goal of the following methods is to recover the blue boundary line. | ||
|
|
||
| * Broad Initial Sampling | ||
|
|
||
| #+BEGIN_SRC jupyter-python | ||
| bs = bsam([[1.5, 1.5]], polygon, boundmin=0, boundmax=3) | ||
| bs.first_sample(ni=16) | ||
| bs.iter_sample(min_points=2, samples=4) | ||
| bs.plot_final(bs.X, bs.cat) | ||
| #+END_SRC | ||
|
|
||
| #+RESULTS: | ||
| :RESULTS: | ||
| [[file:./.ob-jupyter/4ea6a97432b978a6811c907d033b21e510e63977.png]] | ||
| :END: | ||
|
|
||
| To initialize, bsam takes a center point, a polygon shape, and a minimum and maximum bound. The first sample generates a distribution of points around the center point provided. ni is a hyperparameter that determines how many points to initially sample. Other parameters can be found in the broad initial sampling py file in sbs. | ||
|
|
||
| iter_sample performs the simulation to recover the classification boundary within a convergence criteria. The convergence criteria are also hyperparameters that can be specified. | ||
|
|
||
| plot_final plots the true boundary, sample points and their respective classes, predicted boundary, and 95% confidence interval. | ||
|
|
||
| See documentation for hyperparameters. | ||
|
|
||
| * Low Point-Density Sampling | ||
|
|
||
| #+BEGIN_SRC jupyter-python | ||
| ds = dsam([1, 1], polygon, 0, 3) | ||
| ds.first_sample(dis=0.2) | ||
| ds.iter_sample(tol=0.01, n=4, rad_den=0.2, conv_trials=4, | ||
| min_points=1, domain_step=0.15) | ||
| ds.plot_final(ds.X, ds.cat) | ||
| #+END_SRC | ||
|
|
||
| #+RESULTS: | ||
| :RESULTS: | ||
| [[file:./.ob-jupyter/1c305e5cca22fa5631e6b168481bcbe1b23fdff9.png]] | ||
| :END: | ||
|
|
||
|
|
||
| To initialize, dsam takes a center point, a polygon shape, and a minimum and maximum bound. The first sample generates a distribution of points around the center point provided. dis is a hyperparameter that determines the initial radius for sample point distribution. Other parameters can be found in the broad initial sampling py file in sbs. | ||
|
|
||
| iter_sample performs the simulation to recover the classification boundary within a convergence criteria. The convergence criteria are also hyperparameters that can be specified. | ||
|
|
||
| plot_final plots the true boundary, sample points and their respective classes, predicted boundary, and 95% confidence interval. | ||
|
|
||
| See documentation for hyperparameters. | ||
|
|
||
| * Path Tracing | ||
| #+BEGIN_SRC jupyter-python | ||
| ps = psam([[1, 1]], polygon, 0, 3) | ||
| ps.first_sample(dis=0.3, seed=10) | ||
| ps.iter_sample(step=ps.scale*2, dis=0.1, ran_sam=False, | ||
| conv_num=3, centol = ps.scale*3, seed=10) | ||
| ps.plot_final(ps.X, ps.cat) | ||
| #+END_SRC | ||
|
|
||
| #+RESULTS: | ||
| :RESULTS: | ||
| [[file:./.ob-jupyter/6d0fe62410ee12fd1e3a2b58b759e8c008c642db.png]] | ||
| :END: | ||
|
|
||
| To initialize, psam takes a center point, a polygon shape, and a minimum and maximum bound. The first sample generates a distribution of points around the center point provided. dis is a hyperparameter that determines the initial radius for sample point distribution. Other parameters can be found in the broad initial sampling py file in sbs. | ||
|
|
||
| iter_sample performs the simulation to recover the classification boundary within a convergence criteria. The convergence criteria are also hyperparameters that can be specified. | ||
|
|
||
| plot_final plots the true boundary, sample points and their respective classes, predicted boundary, and 95% confidence interval. | ||
|
|
||
| See documentation for hyperparameters. |