How to write the basic code of Java linear regression

This article mainly introduces "how to write the basic code of Java linear regression". In the daily operation, I believe that many people have doubts about how to write the basic code of Java linear regression. The editor consulted all kinds of materials and sorted out simple and easy-to-use operation methods. I hope it will be helpful to answer the doubt of "how to write the basic code of Java linear regression". Next, please follow the editor to study!

# Use linear model to model this data.from sklearn.linear_model import LinearRegressionimport numpy as nplr=LinearRegression () lr.fit (pga.distance [:, np.newaxis], pga ['accuracy']) # Another way is using pga [[' distance']] theta0=lr.intercept_theta1=lr.coef_print (theta0) print (theta1) # calculating cost-function for each theta1# calculates the average cumulative error def cost Theta1): Juni0 for i in range (len (x)): mse= (x [I] * theta1+theta0-y [I]) * * 2 J+=mse return J / (2*len (x)) theta0=100theta1s = np.linspace (- 3 for theta1 in theta1s 2197) costs= [] for theta1 in theta1s: costs.append (cost (pga ['distance'], pga [' accuracy'], theta0,theta1)) plt.plot (theta1s,costs) plt.show () print (pga.distance) # adjust thetadef partial_cost_theta0 Theta1): # when our model is a linear fitting function: y=theta1*x + theta0 Instead of sigmoid function, when nonlinear, we can directly use sigmoid # to operate the whole x series directly, saving a calculation, and finally finding the sum re-average h=theta1*x+theta0 diff= (hmury) partial=diff.sum () / len (diff) return partialpartial0=partial_cost_theta0 (pga.distance,pga.accuracy,1,1) def partial_cost_theta1 (xmemy theta0theta1): # our model is a linear fitting function: y=theta1*x + theta0, not a sigmoid function When non-linear, we can use sigmoid h=theta1*x+theta0 diff= (Hmury) * x partial=diff.sum () / len (diff) return partialpartial1=partial_cost_theta1 (pga.distance,pga.accuracy,0,5) print (partial0) print (partial1) def gradient_descent (x journal yjournal alpha 0.1): # set the default parameter # calculate the cost # adjust the weight # calculate the error cost Judge whether it converges or reaches the maximum number of iterations most_iterations=1000 convergence_thres=0.000001 c=cost [c] cost_pre=c+convergence_thres+1.0 counter=0 while ((np.abs (c-cost_pre) > convergence_thres) & (counter)