在网上看了一下有相关的例子，不过没有完整源码，运行还会报错，稍微做修改和优化，修复部分已知错误。

特别提醒：仅供娱乐，请勿当真，双色球属于随机数，线性回归无法预测随机数的哦~~

编译运行环境：Python 3.7x

原版源码如下：（URL暂时无法正常使用了，改用升级版）

#!/usr/bin/python
# -*- coding:UTF-8 -*-

# 导入需要的包
import os
import pandas as pd
import re
from urllib import request
from bs4 import BeautifulSoup
from sklearn import datasets, linear_model


def get_http_content(href):
    """读取网页内容"""
    request_headers = {
        'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:81.0) Gecko/20100101 Firefox/81.0'
    }
    req = request.Request(href, headers = request_headers)
    try:
        res = request.urlopen(req)
        return res.read()
    except request.HTTPError as e:
        raise e


def get_page_num(url):
    """获取总页数"""
    num = 0
    content = get_http_content(url)
    soup = BeautifulSoup(content, features='lxml')
    strong = soup.find('td', colspan='7')
    # print(strong)
    if strong:
        result = strong.get_text().split(' ')
        list_num = re.findall("[0-9]{1}", result[1])
        for i in range(len(list_num)):
            num = num * 10 + int(list_num[i])
        return num
    else:
        return 0


def get_history_lotto(page):
    """按页爬取号码数据"""
    url = 'http://kaijiang.zhcw.com/zhcw/html/ssq/list_%d.html'
    href = url % page
    print('load......', page, href)
    content = get_http_content(href)
    soup = BeautifulSoup(content, features='lxml')
    em_list = soup.find_all('em')
    div_list = soup.find_all('td', {'align': 'center'})
    n = 0
    m = 0
    results = []
    group = []
    text = ''
    for div in em_list:
        if n == 0:
            # 日期
            group.append(str(div_list[m+0].get_text()))
            # 期数
            group.append(str(div_list[m+1].get_text()))
        # 号码
        lotto = str(div.get_text())
        group.append(lotto)
        # print(lotto)
        n = n + 1
        # 每7个号码组成一组
        if n == 7:
            group.append(text + lotto)
            # 将每组结果汇总到结果列表
            results.append(group)
            group = []
            text = ''
            n = 0  # 重置计数
            m += + 5  # 跳过5列
        else:
            text += lotto + ','
    return results


def get_history_result(out_file):
    """爬取历史数据"""
    url = 'http://kaijiang.zhcw.com/zhcw/html/ssq/list_%d.html'
    # 获取页数
    num = get_page_num(url % 1)
    results = []
    # 按页数循环抓取数据
    for i in range(1, num):
        try:
            # 合并分页查询结果
            results += get_history_lotto(i)
        except Exception as e:
            # 爬取失败时，提示失败原因
            print('error......', i, e)
    # 表头
    columns = {
        'date': str,
        'id': str,  # 期数
        'r1': int,  # 红1
        'r2': int,  # 红2
        'r3': int,  # 红3
        'r4': int,  # 红4
        'r5': int,  # 红5
        'r6': int,  # 红6
        'b1': int,  # 蓝1
        'note': str
    }
    # 生成 DataFrame
    df = pd.DataFrame(columns=columns, data=results)
    print('保存结果')
    df.to_csv(out_file, index=False, encoding='utf-8-sig')
    print(df)
    return df


def get_lotto_data(file, lotto, lotto_id):
    """取读历史数据，指定标识"""
    data = pd.read_csv(file)
    data['lotto_id'] = lotto_id
    X = []
    Y = []
    # single_square_feet, single_price_value
    for s, p in zip(data['lotto_id'], data[lotto]):
        X.append([float(s)])
        Y.append(float(p))
    return X,  Y


# 训练线性回归模型
def linear_model_test(X, Y, predict_value):
    regr = linear_model.LinearRegression()
    regr.fit(X, Y)
    predict_outcome = regr.predict(predict_value)
    predictions = {}
    predictions['intercept'] = regr.intercept_
    predictions['coefficient'] = regr.coef_
    predictions['predicted_value'] = predict_outcome
    return predictions


def get_predicted_num(file, lotto, lotto_id, num):
    """使用线性回归推测中奖号码"""
    X, Y = get_lotto_data(file, lotto, lotto_id)
    # print(X)
    # print(Y)
    predict_value = [[51]]
    result = linear_model_test(X, Y, predict_value)
    print("num", num,
          'Intercept value', result['intercept'],
          'Coefficient', result['coefficient'],
          'Predicted value', result['predicted_value'])


if __name__ == '__main__':
    force_refresh = False  # 强制刷新数据
    file_name = './union_lotto_history_result.csv'
    if not os.path.exists(file_name) or force_refresh:
        # 获取历史数据
        get_history_result(file_name)
    # 使用线性回归模型进行预测 (o_o 预测结果纯属娱乐，请勿深信)
    get_predicted_num(file_name, 'r1', 1, 1)  # 预测红1
    get_predicted_num(file_name, 'r2', 2, 2)  # 预测红2
    get_predicted_num(file_name, 'r3', 3, 3)  # 预测红3
    get_predicted_num(file_name, 'r4', 4, 4)  # 预测红4
    get_predicted_num(file_name, 'r5', 5, 5)  # 预测红5
    get_predicted_num(file_name, 'r6', 6, 28)  # 预测红6
    get_predicted_num(file_name, 'b1', 7, 9)  # 预测蓝7

升级版源码如下：

import requests
import json

import pandas as pd
from sklearn import datasets, linear_model


def get_history_lotto():
    """爬取号码数据"""
    url = 'http://www.cwl.gov.cn/cwl_admin/kjxx/findDrawNotice?name=ssq&issueCount=100'
    headers = {
        'Host': 'www.cwl.gov.cn',
        'Referer': 'http://www.cwl.gov.cn/kjxx/ssq/kjgg/',
        'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:81.0) Gecko/20100101 Firefox/81.0'
    }
    columns = {
        'date': str,
        'id': str,  # 期数
        'r1': int,  # 红1
        'r2': int,  # 红2
        'r3': int,  # 红3
        'r4': int,  # 红4
        'r5': int,  # 红5
        'r6': int,  # 红6
        'b1': int,  # 蓝1
        'note': str
    }
    response = requests.get(url=url, headers=headers)
    # print(response.text)
    response_data = json.loads(response.text)
    result = response_data['result']
    results = []
    for i in result:
        dit = {}
        cnt = 1
        dit['id'] = i['code']
        dit['date'] = i['date']
        for j in str(i['red']).split(','):
            dit['r' + str(cnt)] = j
            cnt += 1
        dit['b1'] = i['blue']
        dit['note'] = i['red'] + ',' + i['blue']
        # print(dit)
        results.append(dit)
    return pd.DataFrame(columns=columns, data=results)


def get_lotto_data(data, lotto, lotto_id):
    """取读历史数据，指定标识"""
    data['lotto_id'] = lotto_id
    X = []
    Y = []
    # single_square_feet, single_price_value
    for s, p in zip(data['lotto_id'], data[lotto]):
        X.append([float(s)])
        Y.append(float(p))
    return X, Y


# 训练线性回归模型
def linear_model_test(X, Y, predict_value):
    regr = linear_model.LinearRegression()
    regr.fit(X, Y)
    predict_outcome = regr.predict(predict_value)
    predictions = {}
    predictions['intercept'] = regr.intercept_
    predictions['coefficient'] = regr.coef_
    predictions['predicted_value'] = predict_outcome
    return predictions


def get_predicted_num(file, lotto, lotto_id, num):
    """使用线性回归推测中奖号码"""
    X, Y = get_lotto_data(file, lotto, lotto_id)
    # print(X)
    # print(Y)
    predict_value = [[33]]
    result = linear_model_test(X, Y, predict_value)
    print("num", num,
          'Intercept value', result['intercept'],
          'Coefficient', result['coefficient'],
          'Predicted value', result['predicted_value'])


if __name__ == '__main__':
    force_refresh = False  # 强制刷新数据
    file_name = './data/union_lotto_history_result.csv'
    df = get_history_lotto()
    df.to_csv(file_name, encoding='utf-8-sig', index=False)
    df = df.sample(100)
    # 使用线性回归模型进行预测 (o_o 预测结果纯属娱乐，请勿深信)
    get_predicted_num(df, 'r1', 1, 5)  # 预测红1
    get_predicted_num(df, 'r2', 2, 10)  # 预测红2
    get_predicted_num(df, 'r3', 3, 15)  # 预测红3
    get_predicted_num(df, 'r4', 4, 20)  # 预测红4
    get_predicted_num(df, 'r5', 5, 24)  # 预测红5
    get_predicted_num(df, 'r6', 6, 29)  # 预测红6
    get_predicted_num(df, 'b1', 7, 9)  # 预测蓝7

测试结果：

num 1 Intercept value 4.73266129032258 Coefficient [0.] Predicted value [4.73266129]
num 2 Intercept value 9.621774193548386 Coefficient [0.] Predicted value [9.62177419]
num 3 Intercept value 14.393145161290322 Coefficient [0.] Predicted value [14.39314516]
num 4 Intercept value 19.2625 Coefficient [0.] Predicted value [19.2625]
num 5 Intercept value 24.161290322580644 Coefficient [0.] Predicted value [24.16129032]
num 28 Intercept value 28.938709677419354 Coefficient [0.] Predicted value [28.93870968]
num 9 Intercept value 8.578225806451613 Coefficient [0.] Predicted value [8.57822581]