A better prediction plan for stock price is for reference only.

1. Guide package

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import keras

2. Read data

data = pd.read_csv('000300.SH.csv')
shsz=data.rename(columns={'收盘价(元)':'Close',"日期":"Date"})

3. Mean prediction method

split_p=int(len(shsz)*0.7)
train = shsz[:split_p]
valid = shsz[split_p:]
preds = []
for i in range(0,1371):
    a = train['closing'][len(train)-1371+i:].sum() + sum(preds)
    b = a/1371
    preds.append(b)
valid['Predictions'] = 0
valid['Predictions'] = preds
plt.plot(train['closing'])
plt.plot(valid[['closing', 'Predictions']])

4. Linear regression prediction method

#linear
split_p=int(len(shsz)*0.7)
train = shsz_robot[:split_p]
valid = shsz_robot[split_p:]

x_train = train.drop(['closing','代码','成交量(股)','简称'], axis=1)
y_train = train['closing']
x_valid = valid.drop(['closing','代码','成交量(股)','简称'], axis=1)
y_valid = valid['closing']
# 实现线性回归
from sklearn.linear_model import LinearRegression
model = LinearRegression()
model.fit(x_train,y_train)
# 做出预测并找到rmse
preds = model.predict(x_valid)
rms = np.sqrt(np.mean(np.power((np.array(y_valid)-np.array(preds)),2)))
# 图表
valid=valid.drop(['代码','成交量(股)','简称'], axis=1)
valid['Predictions'] = 0
valid['Predictions'] = preds
valid.index = shsz_robot[split_p:].index
train.index = shsz_robot[:split_p].index
plt.plot(train['closing'])
plt.plot(valid[['closing', 'Predictions']])

5. k nearest neighbor prediction method

#KNN
from sklearn import neighbors 
from sklearn.model_selection import GridSearchCV 
from sklearn.preprocessing import MinMaxScaler
scaler = MinMaxScaler(feature_range=(0, 1))
x_train_scaled = scaler.fit_transform(x_train)
x_train = pd.DataFrame(x_train_scaled)
x_valid_scaled = scaler.fit_transform(x_valid)
x_valid = pd.DataFrame(x_valid_scaled)# 使用gridsearch查找最佳参数params = {'n_neighbors':[2,3,4,5,6,7,8,9]}knn = neighbors.KNeighborsRegressor()model = GridSearchCV(knn, params, cv=5)
model.fit(x_train,y_train)
preds = model.predict(x_valid)
rmserms = np.sqrt(np.mean(np.power((np.array(y_valid)-np.array(preds)),2)))
rmserms
valid['Predictions'] = 0 
valid['Predictions'] = preds 
plt.plot(train['closing'])
plt.plot(valid[['closing', 'Predictions']]) 

 

6. LSTM prediction method

# 导入所需的库
from sklearn.preprocessing import MinMaxScaler
from keras.models import Sequential
from keras.layers import Dense, Dropout, LSTM

# 创建数据框
data = shsz.sort_index(ascending=True, axis=0)
new_data = pd.DataFrame(index=range(0,len(shsz)),columns=['Date', 'Close'])
for i in range(0,len(data)):
    new_data['Date'][i] = data['Date'][i]
    new_data['Close'][i] = data['Close'][i]

# 设置索引
new_data.index = new_data.Date
new_data.drop('Date', axis=1, inplace=True)

# 创建训练集和验证集
dataset = new_data.values

train = dataset[0:3170,:]
valid = dataset[3170:,:]

# 将数据集转换为x_train和y_train
scaler = MinMaxScaler(feature_range=(0, 1))
scaled_data = scaler.fit_transform(dataset)

x_train, y_train = [], []
for i in range(60,len(train)):
    x_train.append(scaled_data[i-60:i,0])
    y_train.append(scaled_data[i,0])
x_train, y_train = np.array(x_train), np.array(y_train)

x_train = np.reshape(x_train, (x_train.shape[0],x_train.shape[1],1))

train = new_data[:3170]
valid = new_data[3170:]
valid['Predictions'] = closing_price
# plt.plot(train['Close']) 
plt.plot(list(valid['Close']))
plt.plot(list(valid['Predictions']))

# 创建和拟合LSTM网络
model = Sequential()
model.add(LSTM(units=50, return_sequences=True, input_shape=(x_train.shape[1],1)))
model.add(LSTM(units=50))
model.add(Dense(1))

model.compile(loss='mean_squared_error', optimizer='adam')
model.fit(x_train, y_train, epochs=1, batch_size=1, verbose=2)

# 使用过去值来预测246个值
inputs = new_data[len(new_data) - len(valid) - 60:].values
inputs = inputs.reshape(-1,1)
inputs  = scaler.transform(inputs)

X_test = []
for i in range(60,inputs.shape[0]):
    X_test.append(inputs[i-60:i,0])
X_test = np.array(X_test)

X_test = np.reshape(X_test, (X_test.shape[0],X_test.shape[1],1))
closing_price = model.predict(X_test)
closing_price = scaler.inverse_transform(closing_price)

Summarize:

Among the above prediction methods, the more accurate ones are LSTM and K nearest neighbor method. Other methods are relatively rough, but they can also be regarded as a kind of prediction data. However, the above prediction method only represents normal fluctuations, and the futures and stock markets are relatively volatile, and are for reference only.

Python Implementation: Using Machine Learning and Deep Learning to Predict Stock Prices - Zhihu