Method leastSquaresFit

output/java_guava/1.4.17/PairedStats.java:200–216

Returns a linear transformation giving the best fit to the data according to Ordinary Least Squares linear regression of y as a function of x. The count must be greater than one, and either the x or y

()

  public LinearTransformation leastSquaresFit() {
    checkState(count() > 1);
    if (isNaN(sumOfProductsOfDeltas)) {
      return LinearTransformation.forNaN();
    }
    double xSumOfSquaresOfDeltas = xStats.sumOfSquaresOfDeltas();
    if (xSumOfSquaresOfDeltas > 0.0) {
      if (yStats.sumOfSquaresOfDeltas() > 0.0) {
        return LinearTransformation.mapping(xStats.mean(), yStats.mean()).withSlope(sumOfProductsOfDeltas / xSumOfSquaresOfDeltas);
      } else {
        return LinearTransformation.horizontal(yStats.mean());
      }
    } else {
      checkState(yStats.sumOfSquaresOfDeltas() > 0.0);
      return LinearTransformation.vertical(xStats.mean());
    }
  }

  /**
   * {@inheritDoc}