Machine Learning Lab机器学习实验台

Train linear regression, logistic regression and K-means clustering in the browser and see the model plotted live.

在浏览器里直接训练线性回归、逻辑回归与 K-means 聚类，实时画出拟合直线、决策边界与聚类中心。

What is Machine Learning Lab?机器学习实验台 是什么？

A machine-learning lab that trains real models in your browser: ordinary least squares for linear regression, gradient descent for logistic regression, and Lloyd's algorithm for K-means clustering. You pick the dataset and the noise level, hit train, and the model is fitted numerically on the spot — the fitted line, the decision boundary or the cluster centres are drawn on a canvas together with the metrics that matter, such as RMSE, accuracy and iteration count. There is no server call and no model download: the maths is implemented in plain JavaScript, so the whole lab works offline.

一个真正能训练模型的机器学习实验台：线性回归用最小二乘正规方程求解，逻辑回归用梯度下降拟合，K-means 用 Lloyd 算法迭代。你选数据集和噪声强度，点训练，模型当场数值拟合完成——拟合直线、决策边界或聚类中心会画到画布上，并给出真正有意义的指标：RMSE、准确率、迭代次数。全程没有服务器调用，也没有模型下载，数学部分用纯 JavaScript 实现，断网也能用。

How to use machine learning lab如何使用 机器学习实验台

1. Choose an algorithm: linear regression, logistic regression or K-means.
2. Choose a dataset — two Gaussian blobs, a noisy line, or upload your own CSV with x,y columns.
3. Tune sample count, noise level and random seed, then press Generate & train.
4. Read the model parameters and metrics, and export the generated dataset as CSV.

1. 选择算法：线性回归、逻辑回归或 K-means。
2. 选择数据集——双高斯团、带噪声直线，或上传自己的 CSV（x,y 两列）。
3. 调整样本数、噪声与随机种子，点"生成并训练"。
4. 读取模型参数与指标，可把生成的数据集导出为 CSV。

Common uses常见用途

• See how a regression line changes when you add noise to the data
• Understand a decision boundary by watching logistic regression separate two clusters
• Demonstrate K-means without installing Python or Jupyter
• Check whether a classification problem is actually linearly separable

• 观察加入噪声后回归直线如何变化
• 看逻辑回归如何把两团数据分开，直观理解决策边界
• 不装 Python、不装 Jupyter 也能演示 K-means
• 检验一个分类问题是否真的线性可分

FAQ常见问题