线性回归问题和总结

时间:2021-01-12 15:54:41

标签: python pandas plot linear-regression sklearn-pandas

我想创建一个线性回归模型图,显示每年的自行车销量汇总在一个点上,而不是像现在有两个点分开。

这是我的代码:

from sklearn.linear_model import LinearRegression
from sklearn import datasets, linear_model

## Wzrost lub maleje zakup rowerow
## (Purchase of bicycles increases or decreases)
plot1 = df.groupby('Year')['Product_Category'].value_counts().rename('count').reset_index()

x = plot1['Year'].values.reshape(-1, 1)
y = plot1['count'].values.reshape(-1, 1)

# plot #
## linear ##
regr = linear_model.LinearRegression()
regr.fit(x, y)
y_pred = regr.predict(x_test)

#plot#
plt.scatter(x, y,  color='black')
plt.plot(x, y, color='blue', linewidth=3)

这是我的情节:

enter image description here

1 个答案:

答案 0 :(得分:1)

据我从您的示例中了解到,这可能是一个解决方案,将 value_counts 替换为 count

示例数据:

import pandas as pd
import matplotlib.pyplot as plt

df = pd.DataFrame({'Year': [ 2019, 2019, 2020, 2021], 'Product_Category': ['a', 'b', 'c', 'd']})
print(df)
   Year Product_Category
0  2019                a
1  2019                b
2  2020                c
3  2021                d

计数将返回:

plot1 = df.groupby('Year')['Product_Category'].count().rename('count').reset_index()
print(plot1)

  Year  count
0  2019      2
1  2020      1
2  2021      1


plot1 = df.groupby('Year')['Product_Category'].count().rename('count').reset_index()
#x,y#
x = plot1['Year'].values.reshape(-1, 1)
y = plot1['count'].values.reshape(-1, 1)
# plot #

#plot#
plt.scatter(x, y,  color='black')
plt.plot(x, y, color='blue', linewidth=3)

enter image description here