angular-cn/packages/benchpress/src/statistic.ts

/**
 * @license
 * Copyright Google Inc. All Rights Reserved.
 *
 * Use of this source code is governed by an MIT-style license that can be
 * found in the LICENSE file at https://angular.io/license
 */

export class Statistic {
  static calculateCoefficientOfVariation(sample: number[], mean: number) {
    return Statistic.calculateStandardDeviation(sample, mean) / mean * 100;
  }

  static calculateMean(samples: number[]) {
    let total = 0;
    // TODO: use reduce
    samples.forEach(x => total += x);
    return total / samples.length;
  }

  static calculateStandardDeviation(samples: number[], mean: number) {
    let deviation = 0;
    // TODO: use reduce
    samples.forEach(x => deviation += Math.pow(x - mean, 2));
    deviation = deviation / (samples.length);
    deviation = Math.sqrt(deviation);
    return deviation;
  }

  static calculateRegressionSlope(
      xValues: number[], xMean: number, yValues: number[], yMean: number) {
    // See http://en.wikipedia.org/wiki/Simple_linear_regression
    let dividendSum = 0;
    let divisorSum = 0;
    for (let i = 0; i < xValues.length; i++) {
      dividendSum += (xValues[i] - xMean) * (yValues[i] - yMean);
      divisorSum += Math.pow(xValues[i] - xMean, 2);
    }
    return dividendSum / divisorSum;
  }
}