Machine Learning with Swift

Implementing multiple linear regression in Swift

The MultipleLinearRegression class contains a vector of weights, and staff for data normalization:

class MultipleLinearRegression { 
public var weights: [Double]! 
public init() {} 
public var normalization = false 
public var xMeanVec = [Double]() 
public var xStdVec = [Double]() 
public var yMean = 0.0 
public var yStd = 0.0 
... 
} 

Hypothesis and prediction:

public func predict(xVec: [Double]) -> Double { if normalization { let input = xVec let differenceVec = vecSubtract([1.0]+input, xMeanVec) let normalizedInputVec = vecDivide(differenceVec, xStdVec) let h = hypothesis(xVec: normalizedInputVec) return h * yStd + yMean } else { return hypothesis(xVec: [1.0]+xVec) } } private func hypothesis(xVec: ...