I´m trying to find a solution to migrate a logistic regression indicator from Tradingview , but not had much success yet.
I have the source code from Tradingview.
I attach the source code.
//@version=3
// Logistic Regression
study(“LogReg”, overlay=true)
//
// Functions
//
dot(v, w, p) => sum(v * w, p)
sigmoid(z) => 1 / (1 + exp(-z))
logistic_regression(X,Y,p,W,lr,steps)=>
w = 0.0, loss = 0.0
for i=0 to steps
hypothesis = sigmoid(dot(X, W, p)) // prediction
loss := -1 / p * (dot(dot(Y, log(hypothesis) + (1 – Y), p), log(1 – hypothesis), p)) // loss function, gradient, training
gradient = 1 / p * (dot(X, hypothesis – Y, p))
w := w – lr * gradient
[loss, sigmoid(dot(X, w, p))] // current loss & prediction
scaleMinimax(X,p,min,max)=>
hi = highest(X, p), lo = lowest(X, p)
(max – min) * (X – lo)/(hi – lo) + min
//
// Inputs
//
X = input(close)
p = input(3)
//
// Main
//
Y = log(abs(pow(X, 2) – 1) + .5) // generate dataset
W = 0.0
[x,y]=logistic_regression(X,Y,p,W,0.001,100)
x_scaled = scaleMinimax(x, 1000, lowest(X, 1000), highest(X, 1000))
y_scaled = scaleMinimax(y, 1000, lowest(X, 1000), highest(X, 1000))
plot(x_scaled, linewidth=3)
plot(y_scaled, linewidth=3, color=lime)
.
