如何为 sklearn.svm.SVC 定义自定义内核函数？

Posted 2023-03-12

【中文标题】如何为 sklearn.svm.SVC 定义自定义内核函数？

【英文标题】：How can I define a custom kernel function for sklearn.svm.SVC?

【发布时间】：2017-03-31 05:38:20

【问题描述】：

我正在尝试使用 scikit-learn 在 Python 中创建一个股票预测系统。这是我的代码：

import numpy as np
import pandas as pd
from sklearn.preprocessing import StandardScaler
from sklearn.metrics import accuracy_score
import matplotlib.pyplot as plt
from sklearn import svm,preprocessing
from sklearn.metrics import precision_recall_fscore_support
import pandas as pd
import time
##import statistics


def my_kernel(X, Y):
    """
    We create a custom kernel:

                 (2  0)
    k(X, Y) = X  (    ) Y.T
                 (0  1)
    """
    M = np.array([[2, 0], [0, 1.0]])
    return np.dot(np.dot(X, M), Y.T)



FEATURES =  ['DE Ratio',
             'Trailing P/E',
             'Price/Sales',
             'Price/Book',
             'Profit Margin',
             'Operating Margin',
             'Return on Assets',
             'Return on Equity',
             'Revenue Per Share',
             'Market Cap',
             'Enterprise Value',
             'Forward P/E',
             'PEG Ratio',
             'Enterprise Value/Revenue',
             'Enterprise Value/EBITDA',
             'Revenue',
             'Gross Profit',
             'EBITDA',
             'Net Income Avl to Common ',
             'Diluted EPS',
             'Earnings Growth',
             'Revenue Growth',
             'Total Cash',
             'Total Cash Per Share',
             'Total Debt',
             'Current Ratio',
             'Book Value Per Share',
             'Cash Flow',
             'Beta',
             'Held by Insiders',
             'Held by Institutions',
             'Shares Short (as of',
             'Short Ratio',
             'Short % of Float',
             'Shares Short (prior ']

def Build_Data_Set():
    data_df = pd.DataFrame.from_csv("key_stats.csv")
    data_df = data_df.reindex(np.random.permutation(data_df.index))
    ##print data_df
    X = np.array(data_df[FEATURES].values)

    y = (data_df["Status"]
         .replace("underperform",0)
         .replace("outperform",1)
         .values.tolist())

    X = preprocessing.scale(X)
    X = StandardScaler().fit_transform(X)
    Z0 = np.array(data_df["stock_p_hancge"])
    Z1 = np.array(data_df["sp500_p_change"])
    return X,y,Z0,Z1


def mykernel(X, Y,gamma=None):

    X, Y = check_pairwise_arrays(X, Y)
    if gamma is None:
        gamma = 1.0 / X.shape[1]

    K = euclidean_distances(X, Y, squared=True)
    k *= -gamma
    np.exp(K, K)    # exponentiate K in-place
    return safe_sparse_dot(X, Y.T, dense_output=True) + k 

size = 2094
invest_amount = 10000
total_invests = 0
if_market = 0
if_strat = 0
X, y , Z0,Z1= Build_Data_Set()
print(len(X))
test_size = len(X) - size -1 

start = time.clock()
clf = svm.SVC(kernel="mykernel")
clf.fit(X[:size],y[:size])

y_pred = clf.predict(X[size+1:])
y_true = y[size+1:]
time_taken = time.clock()-start
print time_taken,"Seconds"

for x in range(1, test_size+1):
    if y_pred[-x] == 1:
        invest_return = invest_amount + (invest_amount * (Z0[-x]/100))
        market_return = invest_amount + (invest_amount * (Z1[-x]/100))
        total_invests += 1
        if_market += market_return
        if_strat += invest_return

print accuracy_score(y_true, y_pred)

print precision_recall_fscore_support(y_true, y_pred, average='macro')

print "Total Trades:", total_invests
print "Ending with Strategy:",if_strat
print "Ending with Market:",if_market

compared = ((if_strat - if_market) / if_market) * 100.0
do_nothing = total_invests * invest_amount

avg_market = ((if_market - do_nothing) / do_nothing) * 100.0
avg_strat = ((if_strat - do_nothing) / do_nothing) * 100.0


print "Compared to market, we earn",str(compared)+"% more" 
print "Average investment return:", str(avg_strat)+"%" 
print "Average market return:", str(avg_market)+"%" 

预定义的内核正在运行，但对于我的自定义内核，我收到了一个错误：

ValueError: 'mykernel' is not in list

根据官方文档，上面的代码似乎应该可以工作。

【问题讨论】：

请始终在您的问题中显示完整的回溯，而不仅仅是最后一行的错误消息。这样可以更轻松地确定问题所在。

【参考方案1】：

您需要将内核函数本身作为kernel= 参数传递，而不仅仅是函数名称，即：

clf = svm.SVC(kernel=mykernel)

而不是

clf = svm.SVC(kernel="mykernel")

【讨论】：

知道如何将超参数传递给我的内核吗？

@tj89 如果您的内核函数除了X 和Y 之外还有其他参数，您可以使用lambda 或functools.partial 绑定额外的参数，然后再将可调用对象传递给svm.SVC ，例如svm.SVC(kernel=functools.partial(mykernel, param=0.1)) 其中mykernel 是一个接受X, Y, param 的函数。