{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "## Session 2: 初步理解机器学习 (Get some intuitions about Machine Learning by creating a linear regression model)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "___" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### What is linear regression? (什么是线性回归)\n", "\n", "1. 数学概率\n", "2. 一种算法" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "___" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### How we supposed to learn it? (我们应该怎样去学习它呢)\n", "\n", "1. Learn by example\n", "2. Learn by experiment" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "___" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### So the first package you should use is `scikit-learn`\n", "\n", "> Scikit-learn is a free software machine learning library for the Python programming language.\n", "\n", " 1. It is easy for use\n", " 2. It has a good community" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "___" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Import your first ML package (导入你的第一个机器学习包)" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [], "source": [ "from sklearn import linear_model" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [], "source": [ "model = linear_model.LinearRegression()" ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "LinearRegression(copy_X=True, fit_intercept=True, n_jobs=1, normalize=False)" ] }, "execution_count": 8, "metadata": {}, "output_type": "execute_result" } ], "source": [ "model.fit([ [0, 0], [1, 1], [2, 2] ], \n", " [ 0, 1, 2 ])" ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "array([ 3., 55.])" ] }, "execution_count": 10, "metadata": {}, "output_type": "execute_result" } ], "source": [ "model.predict([ [3, 3], [55, 55] ])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "___" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### One more example" ] }, { "cell_type": "code", "execution_count": 19, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "array([28.98113208, 37.09433962])" ] }, "execution_count": 19, "metadata": {}, "output_type": "execute_result" } ], "source": [ "from sklearn import linear_model\n", "\n", "model = linear_model.LinearRegression()\n", "model.fit([[1], [2], [3], [4], [7]],\n", " [ 2, 4, 9, 16, 49])\n", "model.predict([[5], [6]])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "___" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### One more example" ] }, { "cell_type": "code", "execution_count": 21, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "array([8.])" ] }, "execution_count": 21, "metadata": {}, "output_type": "execute_result" } ], "source": [ "from sklearn import linear_model\n", "\n", "model = linear_model.LinearRegression()\n", "model.fit([[1,2,3], [2,3,4], [3,10,1], [8,9,4], [7,4,3]],\n", " [ 6, 9, 14, 21, 14])\n", "model.predict([[5, 1, 2]])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "___" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Convert sklearn model to keras\n", "\n", "1. find problem\n", "2. serach that problem\n", "3. find solution" ] }, { "cell_type": "code", "execution_count": 220, "metadata": {}, "outputs": [], "source": [ "from keras.models import Sequential\n", "\n", "model = Sequential()" ] }, { "cell_type": "code", "execution_count": 221, "metadata": {}, "outputs": [], "source": [ "from keras.layers import Dense\n", "\n", "model.add(Dense(units=2, activation='relu', input_dim=2))\n", "model.add(Dense(units=1))\n", "\n", "model.compile(loss='mean_squared_error',\n", " optimizer='sgd',\n", " metrics=['accuracy'])" ] }, { "cell_type": "code", "execution_count": 222, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Epoch 1/10\n", "5/5 [==============================] - 1s 145ms/step - loss: 28.5372 - acc: 0.0000e+00\n", "Epoch 2/10\n", "5/5 [==============================] - 0s 2ms/step - loss: 16.0311 - acc: 0.0000e+00\n", "Epoch 3/10\n", "5/5 [==============================] - 0s 3ms/step - loss: 13.3427 - acc: 0.0000e+00\n", "Epoch 4/10\n", "5/5 [==============================] - 0s 2ms/step - loss: 11.2660 - acc: 0.0000e+00\n", "Epoch 5/10\n", "5/5 [==============================] - 0s 3ms/step - loss: 9.5507 - acc: 0.2000 \n", "Epoch 6/10\n", "5/5 [==============================] - 0s 4ms/step - loss: 8.0622 - acc: 0.2000\n", "Epoch 7/10\n", "5/5 [==============================] - 0s 2ms/step - loss: 6.8147 - acc: 0.2000\n", "Epoch 8/10\n", "5/5 [==============================] - 0s 2ms/step - loss: 5.8422 - acc: 0.4000\n", "Epoch 9/10\n", "5/5 [==============================] - 0s 3ms/step - loss: 5.1508 - acc: 0.4000\n", "Epoch 10/10\n", "5/5 [==============================] - 0s 4ms/step - loss: 4.6979 - acc: 0.4000\n" ] }, { "data": { "text/plain": [ "" ] }, "execution_count": 222, "metadata": {}, "output_type": "execute_result" } ], "source": [ "import numpy\n", "\n", "x = [[1,1], [2,2], [1,2], [4,3], [1,4]]\n", "y = [ [2] , [4] , [3], [7], [5]]\n", "\n", "x = numpy.array(x)\n", "y = numpy.array(y)\n", "\n", "model.fit(x, y, epochs=10, batch_size=1)" ] }, { "cell_type": "code", "execution_count": 223, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "array([[2.5749485]], dtype=float32)" ] }, "execution_count": 223, "metadata": {}, "output_type": "execute_result" } ], "source": [ "inputs = numpy.array([[2,1]])\n", "model.predict(inputs)" ] } ], "metadata": { "kernelspec": { "display_name": "Python 3", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.6.5" } }, "nbformat": 4, "nbformat_minor": 2 }