1-4 时间序列数据建模流程范例 (国内新冠疫情结束时间预测问题)

文章目录

  • 1-4 时间序列数据建模流程范例 (国内新冠疫情结束时间预测问题)
    • 一，准备数据
    • 二，定义模型
    • 三，训练模型
    • 四，评估模型
    • 五，使用模型
    • 六，保存模型

国内的新冠肺炎疫情从发现至今已经持续很久了，这场起源于吃野味的灾难给大家的生活造成了诸多方面的影响。

有的同学是收入上的，有的同学是感情上的，有的同学是心理上的，还有的同学是体重上的。

那么国内的新冠肺炎疫情何时结束呢？什么时候我们才可以重获自由呢？

本篇文章将利用TensorFlow2.0建立时间序列RNN模型，对国内的新冠肺炎疫情结束时间进行预测。

一，准备数据

本文的数据集取自tushare，数据集在本项目的 data目录下。(这里面只取了2020的三个月的数据)

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import tensorflow as tf
from tensorflow.keras import models,layers,losses,metrics,callbacks 
%matplotlib inline
%config InlineBackend.figure_format = 'svg'df = pd.read_csv("./data/covid-19.csv",sep = "\t")
df.plot(x = "date",y = ["confirmed_num","cured_num","dead_num"],figsize=(10,6))
plt.xticks(rotation=60)

dfdata = df.set_index("date")
dfdiff = dfdata.diff(periods=1).dropna()
dfdiff = dfdiff.reset_index("date")dfdiff.plot(x = "date",y = ["confirmed_num","cured_num","dead_num"],figsize=(10,6))
plt.xticks(rotation=60)
dfdiff = dfdiff.drop("date",axis = 1).astype("float32")

#用某日前8天窗口数据作为输入预测该日数据
WINDOW_SIZE = 8def batch_dataset(dataset):dataset_batched = dataset.batch(WINDOW_SIZE,drop_remainder=True)return dataset_batchedds_data = tf.data.Dataset.from_tensor_slices(tf.constant(dfdiff.values,dtype = tf.float32)) \.window(WINDOW_SIZE,shift=1).flat_map(batch_dataset)ds_label = tf.data.Dataset.from_tensor_slices(tf.constant(dfdiff.values[WINDOW_SIZE:],dtype = tf.float32))#数据较小，可以将全部训练数据放入到一个batch中，提升性能
ds_train = tf.data.Dataset.zip((ds_data,ds_label)).batch(38).cache()

二，定义模型

使用Keras接口有以下3种方式构建模型：使用Sequential按层顺序构建模型，使用函数式API构建任意结构模型，继承Model基类构建自定义模型。

此处选择使用函数式API构建任意结构模型。

#考虑到新增确诊，新增治愈，新增死亡人数数据不可能小于0，设计如下结构
class Block(layers.Layer):def __init__(self, **kwargs):super(Block, self).__init__(**kwargs)def call(self, x_input,x):x_out = tf.maximum((1+x)*x_input[:,-1,:],0.0)return x_outdef get_config(self):  config = super(Block, self).get_config()return config
tf.keras.backend.clear_session()
x_input = layers.Input(shape = (None,3),dtype = tf.float32)
x = layers.LSTM(3,return_sequences = True,input_shape=(None,3))(x_input)
x = layers.LSTM(3,return_sequences = True,input_shape=(None,3))(x)
x = layers.LSTM(3,return_sequences = True,input_shape=(None,3))(x)
x = layers.LSTM(3,input_shape=(None,3))(x)
x = layers.Dense(3)(x)#考虑到新增确诊，新增治愈，新增死亡人数数据不可能小于0，设计如下结构
#x = tf.maximum((1+x)*x_input[:,-1,:],0.0)
x = Block()(x_input,x)
model = models.Model(inputs = [x_input],outputs = [x])
model.summary()
Model: "model"
_________________________________________________________________
Layer (type)                 Output Shape              Param #
=================================================================
input_1 (InputLayer)         [(None, None, 3)]         0
_________________________________________________________________
lstm (LSTM)                  (None, None, 3)           84
_________________________________________________________________
lstm_1 (LSTM)                (None, None, 3)           84
_________________________________________________________________
lstm_2 (LSTM)                (None, None, 3)           84
_________________________________________________________________
lstm_3 (LSTM)                (None, 3)                 84
_________________________________________________________________
dense (Dense)                (None, 3)                 12
_________________________________________________________________
block (Block)                (None, 3)                 0
=================================================================
Total params: 348
Trainable params: 348
Non-trainable params: 0
_________________________________________________________________

三，训练模型

训练模型通常有3种方法，内置fit方法，内置train_on_batch方法，以及自定义训练循环。此处我们选择最常用也最简单的内置fit方法。

注：循环神经网络调试较为困难，需要设置多个不同的学习率多次尝试，以取得较好的效果。

#自定义损失函数，考虑平方差和预测目标的比值
class MSPE(losses.Loss):def call(self,y_true,y_pred):err_percent = (y_true - y_pred)**2/(tf.maximum(y_true**2,1e-7))mean_err_percent = tf.reduce_mean(err_percent)return mean_err_percentdef get_config(self):config = super(MSPE, self).get_config()return config
import os
import datetimeoptimizer = tf.keras.optimizers.Adam(learning_rate=0.01)
model.compile(optimizer=optimizer,loss=MSPE(name = "MSPE"))stamp = datetime.datetime.now().strftime("%Y%m%d-%H%M%S")
logdir = os.path.join('data', 'autograph', stamp)## 在 Python3 下建议使用 pathlib 修正各操作系统的路径
# from pathlib import Path
# stamp = datetime.datetime.now().strftime("%Y%m%d-%H%M%S")
# logdir = str(Path('./data/autograph/' + stamp))tb_callback = tf.keras.callbacks.TensorBoard(logdir, histogram_freq=1)
#如果loss在100个epoch后没有提升，学习率减半。
lr_callback = tf.keras.callbacks.ReduceLROnPlateau(monitor="loss",factor = 0.5, patience = 100)
#当loss在200个epoch后没有提升，则提前终止训练。
stop_callback = tf.keras.callbacks.EarlyStopping(monitor = "loss", patience= 200)
callbacks_list = [tb_callback,lr_callback,stop_callback]history = model.fit(ds_train,epochs=500,callbacks = callbacks_list)
Epoch 371/500
1/1 [==============================] - 0s 61ms/step - loss: 0.1184
Epoch 372/500
1/1 [==============================] - 0s 64ms/step - loss: 0.1177
Epoch 373/500
1/1 [==============================] - 0s 56ms/step - loss: 0.1169
Epoch 374/500
1/1 [==============================] - 0s 50ms/step - loss: 0.1161
Epoch 375/500
1/1 [==============================] - 0s 55ms/step - loss: 0.1154
Epoch 376/500
1/1 [==============================] - 0s 55ms/step - loss: 0.1147
Epoch 377/500
1/1 [==============================] - 0s 62ms/step - loss: 0.1140
Epoch 378/500
1/1 [==============================] - 0s 93ms/step - loss: 0.1133
Epoch 379/500
1/1 [==============================] - 0s 85ms/step - loss: 0.1126
Epoch 380/500
1/1 [==============================] - 0s 68ms/step - loss: 0.1119
Epoch 381/500
1/1 [==============================] - 0s 52ms/step - loss: 0.1113
Epoch 382/500
1/1 [==============================] - 0s 54ms/step - loss: 0.1107
Epoch 383/500
1/1 [==============================] - 0s 55ms/step - loss: 0.1100
Epoch 384/500
1/1 [==============================] - 0s 56ms/step - loss: 0.1094
Epoch 385/500
1/1 [==============================] - 0s 54ms/step - loss: 0.1088
Epoch 386/500
1/1 [==============================] - 0s 74ms/step - loss: 0.1082
Epoch 387/500
1/1 [==============================] - 0s 60ms/step - loss: 0.1077
Epoch 388/500
1/1 [==============================] - 0s 52ms/step - loss: 0.1071
Epoch 389/500
1/1 [==============================] - 0s 52ms/step - loss: 0.1066
Epoch 390/500
1/1 [==============================] - 0s 56ms/step - loss: 0.1060
Epoch 391/500
1/1 [==============================] - 0s 61ms/step - loss: 0.1055
Epoch 392/500
1/1 [==============================] - 0s 60ms/step - loss: 0.1050
Epoch 393/500
1/1 [==============================] - 0s 59ms/step - loss: 0.1045
Epoch 394/500
1/1 [==============================] - 0s 65ms/step - loss: 0.1040
Epoch 395/500
1/1 [==============================] - 0s 58ms/step - loss: 0.1035
Epoch 396/500
1/1 [==============================] - 0s 52ms/step - loss: 0.1031
Epoch 397/500
1/1 [==============================] - 0s 58ms/step - loss: 0.1026
Epoch 398/500
1/1 [==============================] - 0s 60ms/step - loss: 0.1022
Epoch 399/500
1/1 [==============================] - 0s 57ms/step - loss: 0.1017
Epoch 400/500
1/1 [==============================] - 0s 63ms/step - loss: 0.1013
Epoch 401/500
1/1 [==============================] - 0s 59ms/step - loss: 0.1009
Epoch 402/500
1/1 [==============================] - 0s 53ms/step - loss: 0.1005
Epoch 403/500
1/1 [==============================] - 0s 56ms/step - loss: 0.1001
Epoch 404/500
1/1 [==============================] - 0s 55ms/step - loss: 0.0997
Epoch 405/500
1/1 [==============================] - 0s 58ms/step - loss: 0.0993
Epoch 406/500
1/1 [==============================] - 0s 53ms/step - loss: 0.0990
Epoch 407/500
1/1 [==============================] - 0s 59ms/step - loss: 0.0986
Epoch 408/500
1/1 [==============================] - 0s 63ms/step - loss: 0.0982
Epoch 409/500
1/1 [==============================] - 0s 67ms/step - loss: 0.0979
Epoch 410/500
1/1 [==============================] - 0s 55ms/step - loss: 0.0976
Epoch 411/500
1/1 [==============================] - 0s 54ms/step - loss: 0.0972
Epoch 412/500
1/1 [==============================] - 0s 55ms/step - loss: 0.0969
Epoch 413/500
1/1 [==============================] - 0s 55ms/step - loss: 0.0966
Epoch 414/500
1/1 [==============================] - 0s 59ms/step - loss: 0.0963
Epoch 415/500
1/1 [==============================] - 0s 60ms/step - loss: 0.0960
Epoch 416/500
1/1 [==============================] - 0s 62ms/step - loss: 0.0957
Epoch 417/500
1/1 [==============================] - 0s 69ms/step - loss: 0.0954
Epoch 418/500
1/1 [==============================] - 0s 60ms/step - loss: 0.0951
Epoch 419/500
1/1 [==============================] - 0s 50ms/step - loss: 0.0948
Epoch 420/500
1/1 [==============================] - 0s 56ms/step - loss: 0.0946
Epoch 421/500
1/1 [==============================] - 0s 57ms/step - loss: 0.0943
Epoch 422/500
1/1 [==============================] - 0s 55ms/step - loss: 0.0941
Epoch 423/500
1/1 [==============================] - 0s 62ms/step - loss: 0.0938
Epoch 424/500
1/1 [==============================] - 0s 60ms/step - loss: 0.0936
Epoch 425/500
1/1 [==============================] - 0s 100ms/step - loss: 0.0933
Epoch 426/500
1/1 [==============================] - 0s 68ms/step - loss: 0.0931
Epoch 427/500
1/1 [==============================] - 0s 60ms/step - loss: 0.0929
Epoch 428/500
1/1 [==============================] - 0s 50ms/step - loss: 0.0926
Epoch 429/500
1/1 [==============================] - 0s 55ms/step - loss: 0.0924
Epoch 430/500
1/1 [==============================] - 0s 57ms/step - loss: 0.0922
Epoch 431/500
1/1 [==============================] - 0s 75ms/step - loss: 0.0920
Epoch 432/500
1/1 [==============================] - 0s 57ms/step - loss: 0.0918
Epoch 433/500
1/1 [==============================] - 0s 77ms/step - loss: 0.0916
Epoch 434/500
1/1 [==============================] - 0s 50ms/step - loss: 0.0914
Epoch 435/500
1/1 [==============================] - 0s 56ms/step - loss: 0.0912
Epoch 436/500
1/1 [==============================] - 0s 60ms/step - loss: 0.0911
Epoch 437/500
1/1 [==============================] - 0s 55ms/step - loss: 0.0909
Epoch 438/500
1/1 [==============================] - 0s 57ms/step - loss: 0.0907
Epoch 439/500
1/1 [==============================] - 0s 59ms/step - loss: 0.0905
Epoch 440/500
1/1 [==============================] - 0s 60ms/step - loss: 0.0904
Epoch 441/500
1/1 [==============================] - 0s 68ms/step - loss: 0.0902
Epoch 442/500
1/1 [==============================] - 0s 73ms/step - loss: 0.0901
Epoch 443/500
1/1 [==============================] - 0s 50ms/step - loss: 0.0899
Epoch 444/500
1/1 [==============================] - 0s 58ms/step - loss: 0.0898
Epoch 445/500
1/1 [==============================] - 0s 56ms/step - loss: 0.0896
Epoch 446/500
1/1 [==============================] - 0s 52ms/step - loss: 0.0895
Epoch 447/500
1/1 [==============================] - 0s 60ms/step - loss: 0.0893
Epoch 448/500
1/1 [==============================] - 0s 64ms/step - loss: 0.0892
Epoch 449/500
1/1 [==============================] - 0s 70ms/step - loss: 0.0891
Epoch 450/500
1/1 [==============================] - 0s 57ms/step - loss: 0.0889
Epoch 451/500
1/1 [==============================] - 0s 53ms/step - loss: 0.0888
Epoch 452/500
1/1 [==============================] - 0s 51ms/step - loss: 0.0887
Epoch 453/500
1/1 [==============================] - 0s 55ms/step - loss: 0.0886
Epoch 454/500
1/1 [==============================] - 0s 58ms/step - loss: 0.0885
Epoch 455/500
1/1 [==============================] - 0s 55ms/step - loss: 0.0883
Epoch 456/500
1/1 [==============================] - 0s 71ms/step - loss: 0.0882
Epoch 457/500
1/1 [==============================] - 0s 50ms/step - loss: 0.0881
Epoch 458/500
1/1 [==============================] - 0s 56ms/step - loss: 0.0880
Epoch 459/500
1/1 [==============================] - 0s 55ms/step - loss: 0.0879
Epoch 460/500
1/1 [==============================] - 0s 57ms/step - loss: 0.0878
Epoch 461/500
1/1 [==============================] - 0s 56ms/step - loss: 0.0878
Epoch 462/500
1/1 [==============================] - 0s 55ms/step - loss: 0.0879
Epoch 463/500
1/1 [==============================] - 0s 60ms/step - loss: 0.0879
Epoch 464/500
1/1 [==============================] - 0s 68ms/step - loss: 0.0888
Epoch 465/500
1/1 [==============================] - 0s 62ms/step - loss: 0.0875
Epoch 466/500
1/1 [==============================] - 0s 55ms/step - loss: 0.0873
Epoch 467/500
1/1 [==============================] - 0s 49ms/step - loss: 0.0872
Epoch 468/500
1/1 [==============================] - 0s 56ms/step - loss: 0.0872
Epoch 469/500
1/1 [==============================] - 0s 55ms/step - loss: 0.0871
Epoch 470/500
1/1 [==============================] - 0s 55ms/step - loss: 0.0871
Epoch 471/500
1/1 [==============================] - 0s 59ms/step - loss: 0.0870
Epoch 472/500
1/1 [==============================] - 0s 68ms/step - loss: 0.0871
Epoch 473/500
1/1 [==============================] - 0s 57ms/step - loss: 0.0869
Epoch 474/500
1/1 [==============================] - 0s 61ms/step - loss: 0.0870
Epoch 475/500
1/1 [==============================] - 0s 47ms/step - loss: 0.0868
Epoch 476/500
1/1 [==============================] - 0s 55ms/step - loss: 0.0868
Epoch 477/500
1/1 [==============================] - 0s 62ms/step - loss: 0.0866
Epoch 478/500
1/1 [==============================] - 0s 58ms/step - loss: 0.0867
Epoch 479/500
1/1 [==============================] - 0s 60ms/step - loss: 0.0865
Epoch 480/500
1/1 [==============================] - 0s 65ms/step - loss: 0.0866
Epoch 481/500
1/1 [==============================] - 0s 58ms/step - loss: 0.0864
Epoch 482/500
1/1 [==============================] - 0s 57ms/step - loss: 0.0865
Epoch 483/500
1/1 [==============================] - 0s 53ms/step - loss: 0.0863
Epoch 484/500
1/1 [==============================] - 0s 56ms/step - loss: 0.0864
Epoch 485/500
1/1 [==============================] - 0s 56ms/step - loss: 0.0862
Epoch 486/500
1/1 [==============================] - 0s 55ms/step - loss: 0.0863
Epoch 487/500
1/1 [==============================] - 0s 52ms/step - loss: 0.0861
Epoch 488/500
1/1 [==============================] - 0s 68ms/step - loss: 0.0862
Epoch 489/500
1/1 [==============================] - 0s 62ms/step - loss: 0.0860
Epoch 490/500
1/1 [==============================] - 0s 57ms/step - loss: 0.0861
Epoch 491/500
1/1 [==============================] - 0s 51ms/step - loss: 0.0859
Epoch 492/500
1/1 [==============================] - 0s 54ms/step - loss: 0.0860
Epoch 493/500
1/1 [==============================] - 0s 51ms/step - loss: 0.0859
Epoch 494/500
1/1 [==============================] - 0s 54ms/step - loss: 0.0860
Epoch 495/500
1/1 [==============================] - 0s 50ms/step - loss: 0.0858
Epoch 496/500
1/1 [==============================] - 0s 69ms/step - loss: 0.0859
Epoch 497/500
1/1 [==============================] - 0s 63ms/step - loss: 0.0857
Epoch 498/500
1/1 [==============================] - 0s 56ms/step - loss: 0.0858
Epoch 499/500
1/1 [==============================] - 0s 54ms/step - loss: 0.0857
Epoch 500/500
1/1 [==============================] - 0s 57ms/step - loss: 0.0858

四，评估模型

评估模型一般要设置验证集或者测试集，由于此例数据较少，我们仅仅可视化损失函数在训练集上的迭代情况。

%matplotlib inline
%config InlineBackend.figure_format = 'svg'import matplotlib.pyplot as pltdef plot_metric(history, metric):train_metrics = history.history[metric]epochs = range(1, len(train_metrics) + 1)plt.plot(epochs, train_metrics, 'bo--')plt.title('Training '+ metric)plt.xlabel("Epochs")plt.ylabel(metric)plt.legend(["train_"+metric])plt.show()
plot_metric(history,"loss")

五，使用模型

此处我们使用模型预测疫情结束时间，即 新增确诊病例为0 的时间。

#使用dfresult记录现有数据以及此后预测的疫情数据
dfresult = dfdiff[["confirmed_num","cured_num","dead_num"]].copy()
dfresult.tail()


#预测此后100天的新增走势,将其结果添加到dfresult中
for i in range(100):arr_predict = model.predict(tf.constant(tf.expand_dims(dfresult.values[-38:,:],axis = 0)))dfpredict = pd.DataFrame(tf.cast(tf.floor(arr_predict),tf.float32).numpy(),columns = dfresult.columns)dfresult = dfresult.append(dfpredict,ignore_index=True)

dfresult.query("confirmed_num==0").head()# 第55天开始新增确诊降为0，第45天对应3月10日，也就是10天后，即预计3月20日新增确诊降为0
# 注：该预测偏乐观


dfresult.query("cured_num==0").head()# 第164天开始新增治愈降为0，第45天对应3月10日，也就是大概4个月后，即7月10日左右全部治愈。
# 注: 该预测偏悲观，并且存在问题，如果将每天新增治愈人数加起来，将超过累计确诊人数。


dfresult.query("dead_num==0").head()# 第60天开始，新增死亡降为0，第45天对应3月10日，也就是大概15天后，即20200325
# 该预测较为合理

六，保存模型

推荐使用TensorFlow原生方式保存模型。

model.save('./data/tf_model_savedmodel', save_format="tf")
print('export saved model.')

model_loaded = tf.keras.models.load_model('./data/tf_model_savedmodel',compile=False)
optimizer = tf.keras.optimizers.Adam(learning_rate=0.001)
model_loaded.compile(optimizer=optimizer,loss=MSPE(name = "MSPE"))
model_loaded.predict(ds_train)

《30天吃掉那只 TensorFlow2.0》 1-4 时间序列数据建模流程范例 (国内新冠疫情结束时间预测问题)相关推荐

  1. 《30天吃掉那只 TensorFlow2.0》 开篇辞（Tensorflow 学习之路）

    <30天吃掉那只 TensorFlow2.0 >开篇辞

  2. 《30天吃掉那只 TensorFlow2.0 》（附下载）

    本书面向读者  本书假定读者有一定的机器学习和深度学习基础,使用过Keras或者Tensorflow1.0或者Pytorch搭建训练过模型. 对于没有任何机器学习和深度学习基础的同学,建议在学习本书时 ...

  3. 重磅开源！《30天吃掉那只 TensorFlow2.0 》（附下载）

    本书面向读者  本书假定读者有一定的机器学习和深度学习基础,使用过Keras或者Tensorflow1.0或者Pytorch搭建训练过模型. 本书写作风格 本书是一本对人类用户极其友善的TensorF ...

  4. 《30天吃掉那只 TensorFlow2.0》 2-2 三种计算图

    2-2 三种计算图 有三种计算图的构建方式:静态计算图,动态计算图,以及Autograph. 在TensorFlow1.0时代,采用的是静态计算图,需要先使用TensorFlow的各种算子创建计算图, ...

  5. titanic数据集_TF2.0结构化数据建模流程范例

    尽管TensorFlow设计上足够灵活,可以用于进行各种复杂的数值计算.但通常人们使用TensorFlow来实现机器学习模型,尤其常用于实现神经网络模型. 从原理上说可以使用张量构建计算图来定义神经网 ...

  6. 30天吃掉那只tensorflow之（2）：使用 cifar10 数据集来训练网络并测试

    文章目录 写在前头 1. Cifar10数据集的介绍.获取 2. 训练集数据可视化 3. 简单数据处理:将标签进行 one-hot 编码转换 4. 构建网络模型 5. 模型训练 6. 训练数据可视化 ...

  7. 只需一台电脑，人人都可用 Python 实现新冠病毒分布实时追踪

    雷锋网 AI 源创评论: 目前,随着新型冠状病毒(COVID-19)的全球范围内蔓延,海外地区现有确诊已经上升至 293683 (数据更新至 2020.03.25 16:54) .为了更好了保证自身健 ...

  8. 《自然语言处理实战入门》 深度学习组件TensorFlow2.0 ---- 文本数据建模流程

    文章大纲 一,准备数据 二,定义模型 三,训练模型 四,评估模型 五,使用模型 六,保存模型 参考文献 文本处理的建模流程,使用清华发布的新闻分类数据集: 中文文本分类数据集THUCNews THUC ...

  9. TensorFlow2.0学习

    文章目录 一.TensorFlow的建模流程 1.1 结构化数据建模流程范例 1.1.1 准备数据 1.1.2 定义模型 1.1.3 训练模型 1.1.4 评估模型 1.1.5 使用模型 1.1.6 ...

最新文章

  1. 如何跟踪log4j漏洞原理及发现绕WAF的tips
  2. 重构世界的真实，6万奖金的超分辨率图像性能挑战赛邀你来战！
  3. 有钱可以多任性？OpenAI提出人肉模型训练，文本摘要全面超越人类表现！
  4. IDEA找到所有的断点
  5. 怎么调用anaconda安装的pip_对于学python的小伙伴来到底该不该安装andconda
  6. mongodb排序_MongoDB排序
  7. python---set集合
  8. 华为交换机通用配置方式方法
  9. DataFrame的ix方法
  10. Analog Trunk Features （模拟中继特征）
  11. 统计信息：SQL执行优化之密钥
  12. Java面试之JVM
  13. 为了搞清深度学习的随机梯度下降我复习了下微积分
  14. Affinity Designer Mac破解版中文永久激活方法
  15. logback 日志脱敏 隐藏PII信息
  16. Console.read()、Console.readline()、Console.readkey()和Console.Write、Console.Writeline()的意思
  17. 比较二进制工具哪家强？
  18. 时间序列——滑动窗口
  19. docker swam单机启动多个springboot实例，log挂载问题
  20. 40_显示来电归属地，监听来电状态的服务

热门文章

  1. 直播--FFPLAY 的原理
  2. Maven下载，安装与配置
  3. 2345看图王阻止文件删除和U盘弹出 - 解决方案
  4. 群“模”乱舞之简单工厂模式
  5. 有关小米6手机更新REC的一些问题记录
  6. C语言中的read()、write()、fputs()、fgets()
  7. 在不损坏C盘的情况下为C盘扩容，适用于Win
  8. 五年连续亏损42亿，蘑菇如何渡劫？
  9. 高效能程序员的修炼 --读书笔记
  10. linux如何设置低电量关机,ubuntu – 当笔记本电脑达到低电量阈值时,如何使笔记本电脑进入睡眠状态？...