小编far*_*mer的帖子

这篇文章是关于LogisticRegressionCV，GridSearchCV和cross_val_score之间的区别。请考虑以下设置：

import numpy as np
from sklearn.datasets import load_digits
from sklearn.linear_model import LogisticRegression, LogisticRegressionCV
from sklearn.model_selection import train_test_split, GridSearchCV, \
     StratifiedKFold, cross_val_score
from sklearn.metrics import confusion_matrix

read = load_digits()
X, y = read.data, read.target
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.3)

在惩罚逻辑回归中，我们需要设置控制正则化的参数C。scikit-learn中有3种通过交叉验证找到最佳C的方法。

Logistic回归

clf = LogisticRegressionCV (Cs = 10, penalty = "l1",
    solver = "saga", scoring = "f1_macro")
clf.fit(X_train, y_train)
confusion_matrix(y_test, clf.predict(X_test))

旁注：文档指出，SAGA和LIBLINEAR是L1惩罚的唯一优化器，而SAGA对于大型数据集则更快。不幸的是，热启动仅适用于Newton-CG和LBFGS。

GridSearchCV

clf = LogisticRegression (penalty = "l1", solver = "saga", warm_start = True)
clf …

我需要使用 选择数据框的一半groupby，其中每个组的大小未知并且可能因组而异。例如：

       index  summary  participant_id
0     130599     17.0              13
1     130601     18.0              13
2     130603     16.0              13
3     130605     15.0              13
4     130607     15.0              13
5     130609     16.0              13
6     130611     17.0              13
7     130613     15.0              13
8     130615     17.0              13
9     130617     17.0              13
10     86789     12.0              14
11     86791      8.0              14
12     86793     21.0              14
13     86795     19.0              14
14     86797     20.0              14
15     86799      9.0              14
16     86801     10.0              14
20    107370      1.0              15
21 …

这个问题在stackoverflow.com/q/2391679上进行

功能的一个典型例子virtual是

class Shape
{
public:
    virtual string draw() = 0;
};

class Circle : public Shape
{
public:
    string draw() { return "Round"; }
};

class Rectangle : public Shape
{
public:
    string draw() { return "Flat"; }
};

void print (Shape& obj)
{
    cout << obj.draw();
}

但是,我们可以auto在C++ 14中传递一个参数

class Circle
{
public:
    string draw() { return "Round"; }
};

class Rectangle
{
public:
    string draw() { return "Flat"; }
};

void print …

关于stackoverflow的类似问题的答案建议更改实例SVR（）中的参数值，但我不知道如何处理它们。

这是我正在使用的代码：

import json
import numpy as np
from sklearn.svm import SVR

f = open('training_data.txt', 'r')
data = json.loads(f.read())
f.close()

f = open('predict_py.txt', 'r')
data1 = json.loads(f.read())
f.close()

features = []
response = []
predict = []

for row in data:
    a = [
        row['star_power'],
        row['view_count'],
        row['like_count'],
        row['dislike_count'],
        row['sentiment_score'],
        row['holidays'],
        row['clashes'],
    ]
    features.append(a)
    response.append(row['collection'])

for row in data1:
    a = [
        row['star_power'],
        row['view_count'],
        row['like_count'],
        row['dislike_count'],
        row['sentiment_score'],
        row['holidays'],
        row['clashes'],
    ]
    predict.append(a)

X = np.array(features).astype(float)
Y = np.array(response).astype(float)
predict = np.array(predict).astype(float)

svm …