A Bitwise Design and Implementation for Privacy-Preserving Data Mining: From Atomic Operations to Advanced Algorithms

Algorithm 5

The algorithm of logistic regression.

	Input: training data
	Output: parameter θ
(1)	Set parameter θ to 0
(2)	Assign learning rate α and iteration number p, respectively
(3)	Calculate partial derivative of cost function for the training data
(4)	Multiply by the previous outcome
(5)	Update θ by the result of step 4
(6)	Repeat steps 3 to 5 for times to obtain θ