philiprehberger-math_kit

Statistics, regression, interpolation, rounding modes, and moving averages for Ruby

Requirements

Ruby >= 3.1

Installation

Add to your Gemfile:

gem "philiprehberger-math_kit"

Or install directly:

gem install philiprehberger-math_kit

Usage

require "philiprehberger/math_kit"

Statistics

Philiprehberger::MathKit::Stats.mean([1, 2, 3, 4, 5])       # => 3.0
Philiprehberger::MathKit::Stats.median([3, 1, 4, 1, 5])      # => 3.0
Philiprehberger::MathKit::Stats.mode([1, 2, 2, 3])            # => [2]
Philiprehberger::MathKit::Stats.variance([2, 4, 4, 4, 5, 5, 7, 9])  # => 4.0
Philiprehberger::MathKit::Stats.stddev([2, 4, 4, 4, 5, 5, 7, 9])    # => 2.0
Philiprehberger::MathKit::Stats.percentile([1, 2, 3, 4, 5], 50)      # => 3.0
Philiprehberger::MathKit::Stats.sum([1, 2, 3])                # => 6
Philiprehberger::MathKit::Stats.range([1, 5, 3, 9, 2])        # => 8

Skewness and Kurtosis

Philiprehberger::MathKit::Stats.skewness([1, 1, 1, 2, 5, 10])  # => positive (right-skewed)
Philiprehberger::MathKit::Stats.kurtosis([2, 3, 4, 5, 6, 7])   # => near 0 (normal-like)

Confidence Intervals

data = [10, 12, 14, 16, 18]
lower, upper = Philiprehberger::MathKit::Stats.confidence_interval(data, level: 0.95)
# => [9.87, 18.13] (approximate)

Correlation and Covariance

Philiprehberger::MathKit::Stats.correlation([1, 2, 3, 4], [2, 4, 6, 8])  # => 1.0
Philiprehberger::MathKit::Stats.covariance([1, 2, 3, 4], [2, 4, 6, 8])   # => 3.333...

Data Normalization

Philiprehberger::MathKit::Stats.normalize([10, 20, 30])       # => [0.0, 0.5, 1.0]
Philiprehberger::MathKit::Stats.standardize([10, 20, 30, 40]) # => z-scores (mean=0, stddev=1)

Robust Statistics

Philiprehberger::MathKit::Stats.median_absolute_deviation([1, 1, 2, 2, 4, 6, 9])  # => 1.0
Philiprehberger::MathKit::Stats.trimmed_mean([1, 2, 3, 4, 100], trim: 0.2)        # => 3.0
Philiprehberger::MathKit::Stats.winsorized_mean([1, 2, 3, 4, 100], trim: 0.2)     # => less affected by outlier

Linear Regression

result = Philiprehberger::MathKit::Regression.linear([1, 2, 3, 4], [2, 4, 6, 8])
result.slope      # => 2.0
result.intercept  # => 0.0
result.r_squared  # => 1.0
result.predict(5) # => 10.0

Interpolation

points = [[0, 0], [5, 10], [10, 20]]
Philiprehberger::MathKit::Interpolation.linear(points, 2.5)  # => 5.0
Philiprehberger::MathKit::Interpolation.linear(points, 7.5)  # => 15.0

Rounding

Philiprehberger::MathKit::Round.bankers(2.5)                  # => 2.0 (round half to even)
Philiprehberger::MathKit::Round.bankers(3.5)                  # => 4.0
Philiprehberger::MathKit::Round.ceiling(2.1)                  # => 3.0
Philiprehberger::MathKit::Round.floor(2.9)                    # => 2.0
Philiprehberger::MathKit::Round.truncate(2.9)                 # => 2.0
Philiprehberger::MathKit::Round.bankers(2.55, precision: 1)   # => 2.6

Moving Averages

Philiprehberger::MathKit::MovingAverage.simple([1, 2, 3, 4, 5], window: 3)      # => [2.0, 3.0, 4.0]
Philiprehberger::MathKit::MovingAverage.exponential([1, 2, 3, 4, 5], alpha: 0.5) # => [1.0, 1.5, 2.25, 3.125, 4.0625]

API

`Stats`

Method	Description
`.mean(values)`	Arithmetic mean
`.median(values)`	Median (middle value or average of two middle)
`.mode(values)`	Mode(s) as array
`.variance(values, population: true)`	Population or sample variance
`.stddev(values, population: true)`	Standard deviation
`.percentile(values, p)`	Percentile (0-100) with linear interpolation
`.sum(values)`	Sum of values
`.range(values)`	Max - min
`.skewness(values)`	Sample skewness (Fisher-Pearson)
`.kurtosis(values)`	Sample excess kurtosis (Fisher definition)
`.confidence_interval(values, level: 0.95)`	Confidence interval for the mean
`.correlation(xs, ys)`	Pearson correlation coefficient
`.covariance(xs, ys)`	Sample covariance
`.normalize(values)`	Min-max normalization to 0..1
`.standardize(values)`	Z-score standardization (mean=0, stddev=1)
`.median_absolute_deviation(values)`	Median absolute deviation
`.trimmed_mean(values, trim: 0.1)`	Trimmed mean (remove fraction from each end)
`.winsorized_mean(values, trim: 0.1)`	Winsorized mean (clamp extremes)

`Regression`

Method	Description
`.linear(xs, ys)`	Ordinary least squares linear regression
`Result#slope`	Slope of the fitted line
`Result#intercept`	Y-intercept of the fitted line
`Result#r_squared`	Coefficient of determination (0 to 1)
`Result#predict(x)`	Predict y for a given x

`Interpolation`

Method	Description
`.linear(points, x)`	Linear interpolation between sorted points

`Round`

Method	Description
`.bankers(value, precision: 0)`	Banker's rounding (round half to even)
`.ceiling(value, precision: 0)`	Round up
`.floor(value, precision: 0)`	Round down
`.truncate(value, precision: 0)`	Truncate toward zero

`MovingAverage`

Method	Description
`.simple(values, window:)`	Simple moving average
`.exponential(values, alpha:)`	Exponential moving average

Development

bundle install
bundle exec rspec
bundle exec rubocop

Support

If you find this project useful:

⭐ Star the repo

🐛 Report issues

💡 Suggest features

❤️ Sponsor development

🌐 All Open Source Projects

💻 GitHub Profile

🔗 LinkedIn Profile

License

MIT