LSTM预测客流量

最新推荐文章于 2025-07-11 16:22:29 发布

月曜日的tawawa

最新推荐文章于 2025-07-11 16:22:29 发布

阅读量9.5k

点赞数 7

CC 4.0 BY-SA版权

分类专栏：深度学习 LSTM 文章标签： Keras LSTM

本文链接：https://blog.youkuaiyun.com/wenshijie123/article/details/88764036

数据来自这里https://www.jianshu.com/p/5d6d5aac4dbd
从上面参考了网络结构，自己没有调优。batch_size 和 epochs 靠经验（ganjue）
考虑到时间有先后没有打乱数据。也可以去下面拿数据
https://github.com/wenshijie/exercise/tree/master/keliuliang

# -*- coding: utf-8 -*-
"""
Created on Thu Mar 21 18:55:41 2019

@author: lenovo
"""

import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
from sklearn.preprocessing import MinMaxScaler
from keras.models import Sequential
from keras.layers import Dense, LSTM
data = pd.read_csv('international-airline-passengers.csv')
data['time'] = pd.to_datetime(data['time'])
data = data.set_index('time')
# 画出趋势图
def get_picture(data = data):
    data['passengers'].plot()
    plt.figure(figsize=(100,50))
    plt.show()
# 转化序列
def processing(data=data,long=11):
    """
    依次转化为11列
    """
    data['passengers'] = data['passengers'].astype(float)
    sample = len(data) - long + 1
    print('得到{}个样本'.format(sample))
    data_sample = []
    for i in range(sample):
        data_s