plt绘制实测值与预测值的散点图

# -*- coding: utf-8 -*-
import numpy as np
import matplotlib.pyplot as plt
from matplotlib.pyplot import MultipleLocator

"""
输出观测值和模型预测值之间的拟合曲线,即拟合模型
决定系数R2，有拟合曲线公式，有1:1线

注意：R2、RMSE是预测和实测计算的，即原来的反演或预测模型的值。拟合曲线模型是预测和实测重新计算的拟合曲线。
"""
# R2计算
def R2(target, predict):
    SSE = sum((target - predict) ** 2)
    SST = sum((target - (target).mean()) ** 2)
    return 1 - (SSE / SST)

if __name__ == '__main__':
    # x: 是观测值; y: 是模型预测值
    x = np.random.random((100))
    y = (x+np.sin(np.random.random((100))))/2

    fig, ax = plt.subplots(figsize=(3, 3), dpi=100)
    # 绘制1:1对角线，linewidth线的粗细，ls线的格式，c线的颜色，
    ax.plot((0, 1), (0, 1), linewidth=1, transform=ax.transAxes, ls='--', c='k', label="1:1 line", alpha=0.5)
    # 绘制点，'o'点的形状，点的颜色，markersize点的大小
    ax.plot(x, y, 'o', c='black', markersize=5)

    # polyfit(x, y, 1)，1代表线性拟合
    # parameter返回的是线性拟合线的斜率和截距
    parameter = np.polyfit(x, y, 1)
    f = np.poly1d(parameter)
    ax.plot(x, f(x), 'r-', lw=1)

    # 计算决定系数R
    r2 = R2(x,y)
    print(r2)

    # 那个框框的设置
    bbox = dict(boxstyle="round", fc='1', alpha=0.)
    bbox = bbox
    # 在图上安放R2和拟合曲线公式，0.05和0.87是位置偏移量，自己调试
    plt.text(0.05, 0.87, "$R^2=%.2f$\n$y=%.2fx+%.2f$" % ((r2), parameter[0], parameter[1]),
             transform=ax.transAxes, size=7, bbox=bbox)

    # 横轴的设置
    ax.set_xlabel('Measured values($g\cdot kg^{-1}$)', fontsize=7)
    ax.set_ylabel("Predicted values($g\cdot kg^{-1}$)", fontsize=7)

    # 设置图片title
    ax.tick_params(labelsize=7)
    ax.set_title("pre", fontsize=7)

    x_major_locator = MultipleLocator(0.1)
    ax.xaxis.set_major_locator(x_major_locator)
    y_major_locator = MultipleLocator(0.1)
    ax.yaxis.set_major_locator(y_major_locator)
    # 坐标轴
    ax.set(xlim=(-0.05, 1.05), ylim=(-0.05, 1.05))

    # plt.savefig("out.png", bbox_inches='tight')
    plt.show()

在给出其他方法计算截距和斜率的方法：

import numpy as np
from scipy.stats import linregress
import statsmodels.api as sm

# Example usage
x = np.array([1, 2, 3, 4, 5])
y = np.array([2, 4, 6, 8, 10])

# 第一种
def calculate_r_squared(x, y):
    '''Calculates the coefficient of determination (R-squared) and tests for significant correlation between two datasets.
    Parameters:
    - x: the first dataset
    - y: the second dataset
    '''

    # Calculate linear regression. slope: 斜率，intercept: 截距
    slope, intercept, r_value, p_value, std_err = linregress(x, y)

    # Calculate coefficient of determination (R-squared)(决定系数，但是这个决定系数是基于linregress回归的结果的)
    r_squared = r_value**2

    # Check for significant correlation(同理,基于linregress回归的结果的)
    if p_value < 0.05:
        result = "Significant correlation"
    else:
        result = "Non-significant correlation"

    return result, r_squared, p_value

result, r_squared, p_value = calculate_r_squared(x, y)
print(result)
print("Coefficient of determination (R-squared):", r_squared)
print("P-value:", p_value)

# 第二种
def calculate_r_squared_sm(x,y):
    # 添加截距项
    x = sm.add_constant(x)

    # 创建线性回归模型
    model = sm.OLS(y, x)

    # 拟合数据
    results = model.fit()

    # 提取决定系数(决定系数，但是这个决定系数是基于OLS回归的结果的)
    r_squared = results.rsquared

    # 计算相关性显著性(同理, 基于OLS回归的结果的)
    p_value = results.f_pvalue

    # 获取截距项和系数
    intercept = results.params[0]  # 截距项
    coefficient = results.params[1]  # 系数

    return r_squared,p_value, [intercept,coefficient]

print(calculate_r_squared_sm(x,y))