参照教材示例，按类别为2~20，按照10折交叉验证，计算准确度，精确度、召回率和F1值，并绘图

由于缺乏具体数据，我将使用sklearn中的鸢尾花数据集来进行示例。

首先，导入所需的库和数据集：

import numpy as np
import matplotlib.pyplot as plt
from sklearn.datasets import load_iris
from sklearn.model_selection import cross_val_score, KFold
from sklearn.linear_model import LogisticRegression

iris = load_iris()
X = iris.data
y = iris.target

然后，定义10折交叉验证和逻辑回归模型：

kf = KFold(n_splits=10, shuffle=True, random_state=42)
lr = LogisticRegression(max_iter=10000)

接下来，按类别进行交叉验证，并计算准确度、精确度、召回率和F1值：

acc_scores = []
prec_scores = []
rec_scores = []
f1_scores = []

for i in range(2, 21):
    y_binary = np.where(y == i, 1, 0)
    acc = cross_val_score(lr, X, y_binary, cv=kf, scoring='accuracy')
    prec = cross_val_score(lr, X, y_binary, cv=kf, scoring='precision')
    rec = cross_val_score(lr, X, y_binary, cv=kf, scoring='recall')
    f1 = cross_val_score(lr, X, y_binary, cv=kf, scoring='f1')
    
    acc_scores.append(np.mean(acc))
    prec_scores.append(np.mean(prec))
    rec_scores.append(np.mean(rec))
    f1_scores.append(np.mean(f1))

最后，绘制相关图表：

x = range(2, 21)

plt.plot(x, acc_scores, label='Accuracy')
plt.plot(x, prec_scores, label='Precision')
plt.plot(x, rec_scores, label='Recall')
plt.plot(x, f1_scores, label='F1')

plt.legend()
plt.xlabel('Class')
plt.ylabel('Score')
plt.title('Evaluation Metrics by Class')
plt.show()

输出的图表如下所示：