Published
Edited
Oct 25, 2021
Importers
Linear and Logistic Regression
To import:
javascript
import {logistic_regression, linear_regression} from "@herbps10/regression"
bernoulli_log_likelihood = (data, x, y, beta) => {
let ll = 0;
for(let i = 0; i < data.length; i++) {
ll += dbernoulli_log(data[i][y], inv_logit(dot(beta, data[i][x])));
}
return ll;
}
bernoulli_log_likelihood_prime = (data, x, y, beta) => {
const derivatives = new Array(beta.length).fill(0);
for(let j = 0; j < beta.length; j++) {
for(let i = 0; i < data.length; i++) {
derivatives[j] += (data[i][y] - inv_logit(dot(beta, data[i][x]))) * data[i][x][j];
}
}
return derivatives;
}
squared_error_loss = (data, x, y, beta) => {
let loss = 0;
for(let i = 0; i < data.length; i++) {
loss += -Math.pow(data[i][y] - dot(beta, data[i][x]), 2);
}
return loss;
}
squared_error_loss_prime = (data, x, y, beta) => {
const derivatives = new Array(beta.length).fill(0);
for(let j = 0; j < beta.length; j++) {
for(let i = 0; i < data.length; i++) {
derivatives[j] += 2 * (data[i][y] - dot(beta, data[i][x])) * data[i][x][j];
}
}
return derivatives;
}
logistic_regression = (data, x, y, options) => {
options = Object.assign({
learning_rate: 1 / data.length,
beta0: new Array(data[0][x].length).fill(1),
maxiter: 1000
}, options);
const results = gradient_ascent((beta) => bernoulli_log_likelihood(data, x, y, beta), (beta) => bernoulli_log_likelihood_prime(data, x, y, beta), options.beta0, options.maxiter, options.learning_rate);
const predict = (x) => {
return inv_logit(dot(x, results.x));
}
const predictions = data.map((d) => {
return predict(d[x]);
});
return ({
beta: results.x,
predictions: predictions,
gradient_ascent: results,
predict: predict
});
return results;
}
linear_regression = (data, x, y, options) => {
options = Object.assign({
learning_rate: 1 / data.length,
beta0: new Array(data[0][x].length).fill(1),
maxiter: 1000
}, options);
const results = gradient_ascent((beta) => squared_error_loss(data, x, y, beta), (beta) => squared_error_loss_prime(data, x, y, beta), options.beta0, options.maxiter, options.learning_rate);
const predict = (x) => {
return dot(x, results.x);
};
const predictions = data.map((d) => {
return predict(d[x]);
});
return ({
beta: results.x,
predictions: predictions,
gradient_ascent: results,
predict: predict
});
return results;
}
logistic_regression(data, "x", "y")
linear_regression(data, "x", "y")
One platform to build and deploy the best data apps
Experiment and prototype by building visualizations in live JavaScript notebooks. Collaborate with your team and decide which concepts to build out.
Use Observable Framework to build data apps locally. Use data loaders to build in any language or library, including Python, SQL, and R.
Seamlessly deploy to Observable. Test before you ship, use automatic deploy-on-commit, and ensure your projects are always up-to-date.