22 lines
1.2 KiB
Python
22 lines
1.2 KiB
Python
|
import numpy as np
|
|||
|
|
'''
|
|||
|
|
cr. ARNN - Support Information
|
|||
|
|
6.2. Wind speed dataset The wind speed dataset, which is provided by the Japan Meteorological Business Support Center,
|
|||
|
|
contains the wind speed (m/s) time series sampled every ∆𝑡 = 10 minutes between 2010 and 2012 from 𝐷 = 155 wind stations (variables) in Wakkanai, Japan9.
|
|||
|
|
As for the 155 stations, their specific locations (latitude and longitude) can be found in the original dataset file 201606241049longitudelatitude.mat accessible in https://github.com/RPcb/ARNN/tree/master/Data/wind%20speed .
|
|||
|
|
We use 𝑚 = 110 time points as the known series and make predictions on the next 𝐿 − 1 = 45 time points. As shown in Figs. 3a-3b of the main text, the performance of ARNN is better than the other methods.
|
|||
|
|
Besides, utilizing this dataset, we tested the robustness of ARNN with different prediction steps in Figs. 3c-3e of the main text, which proved the effectiveness of ARNN in any time region.
|
|||
|
|
'''
|
|||
|
|
# shape -> [155, 157819]
|
|||
|
|
data = np.loadtxt('exp/data/scale_windspeed.txt')
|
|||
|
|
print(data.shape)
|
|||
|
|
|
|||
|
|
import matplotlib.pyplot as plt
|
|||
|
|
|
|||
|
|
plt.figure(figsize=(12, 6))
|
|||
|
|
plt.plot(data[0,:])
|
|||
|
|
plt.title("Wind Speed Data")
|
|||
|
|
plt.xlabel("Time (dt = 10 mins)")
|
|||
|
|
plt.ylabel("Wind Speed (m/s)")
|
|||
|
|
plt.grid(True)
|
|||
|
|
plt.show()
|