# Model structure:

The input in the figure is the same matrix as our time series matrix, so look at the figure below;

# Data analysis chart:

## Correlation heat map:

## Data distribution diagram:

## Training result:

# Full code:

```
# pip install numpy==1.20.3 -i https://pypi.tuna.tsinghua.edu.cn/simple
#
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import random
import tensorflow
from matplotlib.ticker import StrMethodFormatter
# from statsmodels.tsa.stattools import adfuller
i
```