Merge pull request #17 from Hi-Folks/features/14-linear-regression

Linear Regression

README.md

@@ -49,6 +49,7 @@ Stat class has methods to calculate an average or typical value from a populatio
 - harmonicMean(): harmonic mean;
 - correlation(): the Pearson’s correlation coefficient for two inputs;
 - covariance(): the sample covariance of two inputs.
+- linearRegression():
 
 #### Stat::mean( array $data )
 Return the sample arithmetic mean of the array _$data_.
@@ -243,6 +244,27 @@ $correlation = Stat::correlation(
 // -1.0
 ```
 
+#### Stat::linearRegression ( array $x , array $y )
+Return the slope and intercept of simple linear regression  parameters estimated using ordinary least squares.
+Simple linear regression describes relationship between an independent variable *$x* and a dependent variable *$y* in terms of linear function.
+
+```php 
+$years = [1971, 1975, 1979, 1982, 1983];
+$films_total = [1, 2, 3, 4, 5] 
+list($slope, $intercept) = Stat::linearRegression(
+    $years,
+    $films_total
+);
+// 0.31
+// -610.18
+```
+What happens in 2022, according to the samples above?
+
+```php
+round($slope * 2022 + $intercept);
+// 17.0
+```
+
 ### Freq class
 With *Statistics* package you can calculate frequency table.
 A frequency table is list the frequency of various outcomes in a sample.

examples/stat_methods.php

@@ -37,6 +37,12 @@
 // 1.25
 $variance = Stat::variance([2.75, 1.75, 1.25, 0.25, 0.5, 1.25, 3.5]);
 // 1.3720238095238095
+list($slope, $intercept) = Stat::linearRegression(
+    [1971, 1975, 1979, 1982, 1983],
+    [1, 2, 3, 4, 5]
+);
+// 0.31
+// -610.18
 
 
 try {

src/Stat.php

@@ -438,4 +438,47 @@ public static function correlation(array $x, array $y): false|float
 
         return $a / $b;
     }
+
+    /**
+     * @param array<int|float> $x
+     * @param array<int|float> $y
+     * @throws InvalidDataInputException if 2 arrays have different size,
+     * or if the length of arrays are < 2, or if the 2 input arrays has not numeric elements,
+     * or if the elements of the array are constants
+     * @return array<int|float>
+     */
+    public static function linearRegression(array $x, array $y): array
+    {
+        $countX = count($x);
+        $countY = count($y);
+        if ($countX != $countY) {
+            throw new InvalidDataInputException(
+                'Linear regression requires that both inputs have same number of data points.'
+            );
+        }
+        if ($countX < 2) {
+            throw new InvalidDataInputException(
+                'Linear regression requires at least two data points.'
+            );
+        }
+        $sumX = array_sum($x);
+        $sumY = array_sum($y);
+        $sumXX = 0;
+        $sumXY = 0;
+
+        foreach ($x as $key => $value) {
+            $sumXY += ($value * $y[$key]);
+            $sumXX += ($value * $value);
+        }
+        $denominator = (($countX * $sumXX) - ($sumX * $sumX));
+        if ($denominator === 0) {
+            throw new InvalidDataInputException(
+                'Linear regression, the inputs is constant.'
+            );
+        }
+        $slope = (($countX * $sumXY) - ($sumX * $sumY)) / $denominator;
+        $intercept = ($sumY - ($slope * $sumX)) / $countX;
+
+        return [$slope, $intercept];
+    }
 }

tests/StatFromCsvTest.php

@@ -0,0 +1,36 @@
+<?php
+
+
+use HiFolks\Statistics\Stat;
+
+it('parse CSV', function () {
+    $row = 0;
+
+    if (($handle = fopen(getcwd()."/tests/data/income.data.csv", "r")) !== false) {
+        $x = [];
+        $y = [];
+        while (($data = fgetcsv($handle, 1000, ",")) !== false) {
+            $num = count($data);
+            expect($num)->toEqual(3);
+            $row++;
+            if ($row === 1) {
+                continue;
+            }
+            $income = floatval($data[1]);
+            $x[] = $income;
+            $happiness = floatval($data[2]);
+            $y[] = $happiness;
+            expect($income)->toBeFloat();
+            expect($income)->toBeGreaterThan(0);
+            expect($happiness)->toBeFloat();
+        }
+        list($slope, $intercept) = Stat::linearRegression($x, $y);
+        expect(round($slope, 5))->toEqual(0.71383);
+        expect(round($intercept, 5))->toEqual(0.20427);
+        //expect(round(Stat::median($x), 5))->toEqual(0);
+
+        fclose($handle);
+    }
+
+    expect($row)->toEqual(499);
+});

tests/StatTest.php

@@ -318,3 +318,50 @@
         )
     )->toThrow(InvalidDataInputException::class);
 });
+
+it('calculates linear regression (static)', function () {
+    list($slope, $intercept) = Stat::linearRegression(
+        [1971, 1975, 1979, 1982, 1983],
+        [1, 2, 3, 4, 5]
+    );
+    expect($slope)->toBeFloat();
+    expect($slope)->toEqual(0.31);
+
+    expect($intercept)->toBeFloat();
+    expect($intercept)->toEqual(-610.18);
+
+    list($slope, $intercept) = Stat::linearRegression(
+        [1971, 1975, 1979, 1982, 1983],
+        [1, 2, 1, 3, 1]
+    );
+    expect($slope)->toBeFloat();
+    expect($slope)->toEqual(0.05);
+
+    expect($intercept)->toBeFloat();
+    expect($intercept)->toEqual(-97.3);
+
+    expect(round($slope * 2019 + $intercept))->toEqual(4);
+});
+
+it('calculates linear regression with not valid input (static)', function () {
+    expect(
+        fn () => Stat::linearRegression(
+            [3],
+            [2]
+        )
+    )->toThrow(InvalidDataInputException::class);
+
+    expect(
+        fn () => Stat::linearRegression(
+            [3,3,3,3],
+            [2,1,1,1,1]
+        )
+    )->toThrow(InvalidDataInputException::class);
+
+    expect(
+        fn () => Stat::linearRegression(
+            [3,3,3,3,3],
+            [1,1,1,1,1]
+        )
+    )->toThrow(InvalidDataInputException::class);
+});