-
Notifications
You must be signed in to change notification settings - Fork 2
/
Fair-Learning-Classifier.Rmd
42 lines (29 loc) · 1.22 KB
/
Fair-Learning-Classifier.Rmd
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
40
41
42
---
title: "Fair classifier Design"
author: Aytijhya Saha, Samahriti Mukherjee, Rishi Dey Chowdhury
date: February 15, 2022
output: powerpoint_presentation
---
```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = FALSE)
```
## Objective
We present here regularization-based approach for reducing bias in
learned classifiers. We apply this approach using various definitions of fairness to establish fairness in the learned classifier and compare our results to that of a classifier without implementing this fairness based regularization.
## Description of the data
We demonstrate the method using Adult Data, downloaded from <https://archive.ics.uci.edu/ml/machine-learning-databases/adult/>. Here,prediction task is to classify whether a person makes over 50K a year. The protected groups are :
race and sex
The unproteected groups are -
age, workclass, fnlwgt, education, education-num, marital-status,occupation, relationship, capital-gain, capital-loss, hours-per-week and native country
## Brief description of the method:
$minimize_{\theta}\enspace [-ll(\theta;S) + \sum_{i=1}^{N} c_i.D_i + ||\theta||_2] $
## Slide with Plot
```{r pressure}
plot(pressure)
```
## Slide with Plot
```{r}
v=c(1,2,3)
x=c(2,7,4)
plot(v,x)
```