compas

<code><a rd-options="" href="/link/compas?package=fairness&version=1.2.0" data-mini-rdoc="fairness::compas">compas</a></code> is a landmark dataset to study algorithmic (un)fairness. This data was used to
predict recidivism (whether a criminal will reoffend or not) in the USA. The tool was meant to overcome
human biases and offer an algorithmic, fair solution to predict recidivism in a diverse population.
However, the algorithm ended up propagating existing social biases and thus, offered an unfair algorithmic
solution to the problem. In this dataset, a model to predict recidivism has already been fit and predicted
probabilities and predicted status (yes/no) for recidivism have been concatenated to the original data.

datasets

Offers calculation, visualization and comparison of algorithmic fairness metrics. Fair machine learning is an emerging topic with the overarching aim to critically assess whether ML algorithms reinforce existing social biases. Unfair algorithms can propagate such biases and produce predictions with a disparate impact on various sensitive groups of individuals (defined by sex, gender, ethnicity, religion, income, socioeconomic status, physical or mental disabilities). Fair algorithms possess the underlying foundation that these groups should be treated similarly or have similar prediction outcomes. The fairness R package offers the calculation and comparisons of commonly and less commonly used fairness metrics in population subgroups. These methods are described by Calders and Verwer (2010) <doi:10.1007/s10618-010-0190-x>, Chouldechova (2017) <doi:10.1089/big.2016.0047>, Feldman et al. (2015) <doi:10.1145/2783258.2783311> , Friedler et al. (2018) <doi:10.1145/3287560.3287589> and Zafar et al. (2017) <doi:10.1145/3038912.3052660>. The package also offers convenient visualizations to help understand fairness metrics.

Nikita Kozodoi

fairness

Algorithmic Fairness Metrics

Tibor V. Varga

compas function

A data frame with 6172 rows and 9 variables:<dl class="dl-horizontal">
 <dt>Two_yr_Recidivism</dt><dd>factor, yes/no for recidivism or no recidivism. This is the outcome or target in this dataset</dd>
 <dt>Number_of_Priors</dt><dd>numeric, number of priors, normalized to mean = 0 and standard deviation = 1</dd>
 <dt>Age_Above_FourtyFive</dt><dd>factor, yes/no for age above 45 years or not</dd>
 <dt>Age_Below_TwentyFive</dt><dd>factor, yes/no for age below 25 years or not</dd>
 <dt>Female</dt><dd>factor, female/male for gender</dd>
 <dt>Misdemeanor</dt><dd>factor, yes/no for having recorded misdemeanor(s) or not</dd>
 <dt>ethnicity</dt><dd>factor, Caucasian, African American, Asian, Hispanic, Native American or Other</dd>
 <dt>probability</dt><dd>numeric, predicted probabilities for recidivism, ranges from 0 to 1</dd>
 <dt>predicted</dt><dd>numeric, predicted values for recidivism, 0/1 for no/yes</dd>
</dl>

Format

<code><a rd-options='' href='compas'>compas</a></code> is a landmark dataset to study algorithmic (un)fairness. This data was used to
predict recidivism (whether a criminal will reoffend or not) in the USA. The tool was meant to overcome
human biases and offer an algorithmic, fair solution to predict recidivism in a diverse population.
However, the algorithm ended up propagating existing social biases and thus, offered an unfair algorithmic
solution to the problem. In this dataset, a model to predict recidivism has already been fit and predicted
probabilities and predicted status (yes/no) for recidivism have been concatenated to the original data.

Modified COMPAS dataset — compas

A data frame with 6172 rows and 9 variables:<dl class='dl-horizontal'>
 <dt>Two_yr_Recidivism</dt><dd>factor, yes/no for recidivism or no recidivism. This is the outcome or target in this dataset</dd>
 <dt>Number_of_Priors</dt><dd>numeric, number of priors, normalized to mean = 0 and standard deviation = 1</dd>
 <dt>Age_Above_FourtyFive</dt><dd>factor, yes/no for age above 45 years or not</dd>
 <dt>Age_Below_TwentyFive</dt><dd>factor, yes/no for age below 25 years or not</dd>
 <dt>Female</dt><dd>factor, female/male for gender</dd>
 <dt>Misdemeanor</dt><dd>factor, yes/no for having recorded misdemeanor(s) or not</dd>
 <dt>ethnicity</dt><dd>factor, Caucasian, African American, Asian, Hispanic, Native American or Other</dd>
 <dt>probability</dt><dd>numeric, predicted probabilities for recidivism, ranges from 0 to 1</dd>
 <dt>predicted</dt><dd>numeric, predicted values for recidivism, 0/1 for no/yes</dd>
</dl>

compas: Modified COMPAS dataset

Description

Usage

Arguments

Format