03-02 擬合

執行結果:

(1)main.py

import numpy as np
import matplotlib.pyplot as plt
import scipy as sp
import math

#----------------------------------------
# 讀取網路流量資料 (時間序列, 流量)
#----------------------------------------
data=np.genfromtxt('webTraffic.csv',  names=['time', 'amount'], delimiter=',')

# 找出趨近測試資料的n階線性方程式系數
fp1=sp.polyfit(data['time'], data['amount'], 1)
fp2=sp.polyfit(data['time'], data['amount'], 2)
fp3=sp.polyfit(data['time'], data['amount'], 3)

# 擬合方程式
f1=sp.poly1d(fp1)
f2=sp.poly1d(fp2)
f3=sp.poly1d(fp3)

# 印出誤差
print(f1.order, '階-', '誤差:', math.log(sp.sum((data['amount']-f1(data['time']))**2)))
print(f2.order, '階-', '誤差:', math.log(sp.sum((data['amount']-f2(data['time']))**2)))
print(f3.order, '階-', '誤差:', math.log(sp.sum((data['amount']-f3(data['time']))**2)))


#---------------------------
# 繪圖
#---------------------------
fig = plt.figure()

# 設定圖標題
plt.title('Web Traffic')

# 設定x軸及y軸標題
plt.xlabel('time')
plt.ylabel('amount')

#傳回1000個(從0到x元素個數), 間隔相同的數
fx=sp.linspace(0, data['time'][-1], 1000)

# 將測試資料畫在圖上
plt.axis([-100, 800, 0, 7500])
plt.plot(data['time'], data['amount'], 'y+')

# 將產生的數一個個代入線性方程式, 並畫在圖上
plt.plot(fx, f1(fx), 'r-', linewidth=4)
plt.plot(fx, f2(fx), 'g-', linewidth=4)
plt.plot(fx, f3(fx), 'b-', linewidth=4)

# 產生左上角的標籤說明
plt.legend(['data','Power of 1', 'Power of 2', 'Power of 3'], loc="upper left")

# 畫出格線
plt.grid()

#---------------------------
# 儲存圖檔
#---------------------------
fig.savefig('graph.png')

(2)webTraffic.csv

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

03-02 擬合.md

03-02 擬合.md

03-02 擬合

執行結果:

(1)main.py

(2)webTraffic.csv

Files

03-02 擬合.md

Latest commit

History

03-02 擬合.md

File metadata and controls

03-02 擬合

執行結果:

(1)main.py

(2)webTraffic.csv