核心代码
def linear_interpolation(x1, y1, x2, y2, x0):
"""
线性插值函数
参数:
x1, y1: 第一个已知数据点的位置和数据值
x2, y2: 第二个已知数据点的位置和数据值
x0: 要估算的位置
返回值:
y0: 在位置 x0 处估算的数据值
"""
y0 = y1 + (x0 - x1) * (y2 - y1) / (x2 - x1)
return y0
# 示例数据
x1, y1 = 0, 10
x2, y2 = 5, 20
x0 = 2
# 使用线性插值估算在位置 x0 处的数据值
y0 = linear_interpolation(x1, y1, x2, y2, x0)
print(f"在位置 {x0} 处的估算数据值为:{y0}")
数据格式:
分析:1.首先判断该数据是否需要插值
2.获取nan的索引位置,保存到list集合中去
3.根据获取的nan的索引位置,得到位置前后的数据,通过线性插值法算出该nan的值。
代码具体实现:
import numpy as np
import pandas as pd
def is_exist_nan(data):
for d in data:
if np.isnan(d):
return True
return False
def get_nan_index_list(data):
index_list = []
for index,d in enumerate(data):
if np.isnan(d):
index_list.append(index)
return index_list
def linear_interpolation(x1, y1, x2, y2, x0):
"""
线性插值函数
参数:
x1, y1: 第一个已知数据点的位置和数据值
x2, y2: 第二个已知数据点的位置和数据值
x0: 要估算的位置
返回值:
y0: 在位置 x0 处估算的数据值
"""
y0 = y1 + (x0 - x1) * (y2 - y1) / (x2 - x1)
return y0
def get_first_data(data):
for index,d in enumerate(data):
if not np.isnan(d):
return d
def get_last_data(data):
count = len(data) -1
for d in data:
if not np.isnan(data[count]):
return data[count]
else:
count=count-1
def digu(x2,data):
if not np.isnan(data[x2]):
return x2,data
else:
x2 = x2 + 1
return digu(x2,data)
def get_new_data(nan_index_list, data):
if nan_index_list[0] == 0:
data[0] = get_first_data(data)
nan_index_list.remove(0)
if len(nan_index_list)>=1:
if nan_index_list[len(nan_index_list)-1] == 26:
data[26] =get_last_data(data)
nan_index_list.remove(26)
if len(nan_index_list) >=1:
for nan_index in nan_index_list:
x1 = nan_index - 1
y1 = data[x1]
x2 = nan_index + 1
x2,data = digu(x2,data)
y2 = data[x2]
x0 = nan_index
y0 = round(linear_interpolation(x1, y1, x2, y2, x0), 4)
data[nan_index] = y0
return data
if __name__ == '__main__':
data1 = [np.nan, -0.3356, -0.3208, -0.3661, 0.2192, np.nan, np.nan, np.nan, -0.3709, -0.3779, 0.026, -0.2601,
np.nan, -0.0238, -0.2241, -0.2105, -0.2623, 0.379, -0.2196, np.nan, -0.0835, 0.2895, 0.0415, -0.2323,
-0.1782, -0.2308, -0.2265]
if is_exist_nan(data1):
print(data1)
nan_index_list = get_nan_index_list(data1)
new_data = get_new_data(nan_index_list,data1)
print(new_data)
运行结果如下:
[nan, -0.3356, -0.3208, -0.3661, 0.2192, nan, nan, nan, -0.3709, -0.3779, 0.026, -0.2601, nan, -0.0238, -0.2241, -0.2105, -0.2623, 0.379, -0.2196, nan, -0.0835, 0.2895, 0.0415, -0.2323, -0.1782, -0.2308, -0.2265]
[-0.3356, -0.3356, -0.3208, -0.3661, 0.2192, 0.0717, -0.0758, -0.2234, -0.3709, -0.3779, 0.026, -0.2601, -0.1419, -0.0238, -0.2241, -0.2105, -0.2623, 0.379, -0.2196, -0.1515, -0.0835, 0.2895, 0.0415, -0.2323, -0.1782, -0.2308, -0.2265]
给大家提供一个思路,具体用的时候,推荐用pandas的interpolate方法实现。
import pandas as pd
if __name__ == '__main__':
# 原始数据,包含缺失值
data = [np.nan, -0.3356, -0.3208, -0.3661, 0.2192, np.nan, np.nan, np.nan, -0.3709, -0.3779, 0.026, -0.2601,
np.nan, -0.0238, -0.2241, -0.2105, -0.2623, 0.379, -0.2196, np.nan, -0.0835, 0.2895, 0.0415, -0.2323,
-0.1782, -0.2308, -0.2265]
# 将数据转换为pandas的Series对象,此时缺失值会自动转换为NaN
data_series = pd.Series(data)
# 执行线性插值,并处理第一个和最后一个NaN
interpolated_data = data_series.interpolate(limit_direction='both')
# 打印插值结果
print(interpolated_data.values)
import numpy as np
import pandas as pd
from scipy.signal import savgol_filter
import matplotlib.pyplot as plt
def get_interpolated_data(data):
# 将数据转换为pandas的Series对象,此时缺失值会自动转换为NaN
data_series = pd.Series(data)
# 执行线性插值,并处理第一个和最后一个NaN
interpolated_data = data_series.interpolate(limit_direction='both').tolist()
interpolated_data = [round(i, 4) for i in interpolated_data] # 保留四位小数
return interpolated_data
def get_sg_data(data, window_size, polyorder):
smoothed_data = savgol_filter(data, window_size, polyorder).tolist()
smoothed_data = [round(i, 4) for i in smoothed_data] # 保留四位小数
return smoothed_data
if __name__ == '__main__':
# 原始数据,包含缺失值
data = [np.nan, -0.3356, -0.3208, -0.3661, 0.2192, np.nan, np.nan, np.nan, -0.3709, -0.3779, 0.026, -0.2601,
np.nan, -0.0238, -0.2241, -0.2105, -0.2623, 0.379, -0.2196, np.nan, -0.0835, 0.2895, 0.0415, -0.2323,
-0.1782, -0.2308, -0.2265]
interpolated_data = get_interpolated_data(data)
sg_data = get_sg_data(interpolated_data,5,2)
sg_data2 = get_sg_data(interpolated_data,9,3)
print(interpolated_data)
print(sg_data)
print(sg_data2)
plt.plot(interpolated_data, label='interpolated_data')
plt.plot(sg_data, label='sg_data window_size=5 polyorder=2')
plt.plot(sg_data2, label='sg_data2 window_size=9 polyorder=3')
plt.xlabel('Time')
plt.ylabel('Value')
plt.title('Line Plot')
plt.legend()
plt.show()