如何从 MultinomialNB

如何解决如何从 MultinomialNB

我正在尝试根据 13 个输入值构建一个公式，以预测事件发生的概率。现在，我有一个大约 150 个案例的电子表格，其中每个案例都有 13 个输入值，如果事件发生则为 1，否则为 0（如果你很好奇，数据来自一组NFL传球尝试。13个输入值是投掷的情况，最终值是是否完成）

我目前有一个 MultinomialNB 函数，可以为事件是否发生创建预测函数，该函数似乎运行良好。我用另外 10 个输入集（具有已知结果）对其进行了测试，其中 8 个是正确的。但是，我能从中得到的只是对进一步数据的预测，而不是实际的拟合函数或概率。理想情况下，我希望函数能够返回导致 1 或 0 的事件概率，即不是返回 [1 0 0]，而是返回类似 [.89 .12 .33] 的内容。

如果这不可行，我希望能够查看拟合函数（以一种有意义的方式）。目前我有一个输出值的 predict_proba(X) 函数，但它们不是我期望的形式。我在下面附上了我的代码和结果输出。

from sklearn.naive_bayes import MultinomialNB
clf = MultinomialNB()
NFL_All2 = pd.read_csv("Downloads/NFL Data noNeg.csv")
Parameters = NFL_All2[['Down (5)','distance (6)','Yards to End Zone (7)','distance of Throw (9)','Direction of Throw (10)','QB Motion (11)','Nearest Defender (QB) (12)','Nearest Defender (R) (13)','Defender Position (R) (14)','Contested Catch (15)','Boundary distance (17)','Throw Angle (18)','Throw Speed (19)']]
X = Parameters
y = NFL_All2['Completion (20)']
#Defines parameters for MultinomialNB() function for the given lines of the input spreadsheet

predictNFL = pd.read_csv('Downloads/NFL Data Test.csv')
predictData = predictNFL[['Down (5)','Throw Speed (19)']]
predictData2 = [predictData.iloc[0,:],predictData.iloc[1,predictData.iloc[2,predictData.iloc[3,predictData.iloc[4,predictData.iloc[5,predictData.iloc[6,predictData.iloc[7,predictData.iloc[8,predictData.iloc[9,:]]
# Defines array of data to be predicted (probably in a super inefficient way)

Fit_form = clf.fit(X,y)
print(clf.predict(predictData2))
print(Fit_form.predict_proba(X))
#Fits the data and uses the MultinomialNB fit to predict the ten sets of data in predictData2

输出：

[1 0 1 0 0 1 1 1 0 1]
[[4.83717085e-01 5.16282915e-01]
 [8.80356674e-06 9.99991196e-01]
 [9.95663712e-01 4.33628791e-03]
 [6.34463299e-04 9.99365537e-01]
 [7.18779043e-01 2.81220957e-01]
 [9.89591627e-01 1.04083729e-02]
 [9.03269745e-01 9.67302551e-02]
 [4.01094501e-06 9.99995989e-01]
 [2.86091192e-05 9.99971391e-01]
 [9.99999828e-01 1.71768141e-07]
 [3.38561700e-05 9.99966144e-01]
 [3.96564681e-03 9.96034353e-01]
 [9.04084399e-07 9.99999096e-01]
 [8.45942184e-02 9.15405782e-01]
 [9.90556990e-01 9.44300971e-03]
 [1.20057914e-02 9.87994209e-01]
 [9.98472893e-01 1.52710737e-03]
 [5.62721669e-05 9.99943728e-01]
 [9.99998829e-01 1.17128318e-06]
 [7.18453246e-01 2.81546754e-01]
 [9.99999973e-01 2.74206146e-08]
 [3.05105703e-03 9.96948943e-01]
 [1.35333047e-01 8.64666953e-01]
 [2.20093004e-02 9.77990700e-01]
 [5.52406770e-05 9.99944759e-01]
 [7.09632524e-05 9.99929037e-01]
 [1.09312672e-01 8.90687328e-01]
 [6.50588141e-03 9.93494119e-01]
 [9.36744920e-01 6.32550802e-02]
 [3.44244945e-02 9.65575506e-01]
 [1.01692664e-05 9.99989831e-01]
 [7.55467677e-01 2.44532323e-01]
 [1.01667198e-04 9.99898333e-01]
 [9.98690292e-01 1.30970751e-03]
 [9.88735018e-01 1.12649821e-02]
 [9.48611434e-01 5.13885657e-02]
 [1.43943805e-01 8.56056195e-01]
 [1.04257155e-02 9.89574285e-01]
 [3.22010113e-08 9.99999968e-01]
 [1.98854933e-03 9.98011451e-01]
 [9.99856611e-01 1.43388739e-04]
 [2.80320654e-02 9.71967935e-01]
 [3.70094316e-06 9.99996299e-01]
 [9.97095258e-01 2.90474203e-03]
 [9.45918901e-01 5.40810990e-02]
 [2.65025337e-04 9.99734975e-01]
 [2.98039566e-05 9.99970196e-01]
 [1.20186799e-02 9.87981320e-01]
 [9.86718422e-01 1.32815778e-02]
 [9.99999893e-01 1.06972161e-07]
 [5.03703129e-06 9.99994963e-01]
 [6.97512459e-03 9.93024875e-01]
 [1.22324349e-01 8.77675651e-01]
 [2.73647237e-01 7.26352763e-01]
 [4.78238640e-01 5.21761360e-01]
 [1.48498135e-01 8.51501865e-01]
 [1.03845708e-04 9.99896154e-01]
 [9.99999985e-01 1.49847549e-08]
 [9.84364206e-01 1.56357942e-02]
 [9.99997825e-01 2.17497383e-06]
 [1.17004882e-03 9.98829951e-01]
 [3.27017095e-01 6.72982905e-01]
 [3.36049995e-03 9.96639500e-01]
 [6.33620065e-01 3.66379935e-01]
 [5.18617803e-02 9.48138220e-01]
 [9.95441768e-01 4.55823188e-03]
 [3.88910822e-03 9.96110892e-01]
 [9.57993464e-03 9.90420065e-01]
 [2.15679691e-04 9.99784320e-01]
 [7.16888650e-01 2.83111350e-01]
 [3.19884319e-04 9.99680116e-01]
 [8.85584817e-06 9.99991144e-01]
 [9.92922292e-01 7.07770837e-03]
 [9.99939553e-01 6.04466563e-05]
 [4.15134849e-03 9.95848652e-01]
 [1.11415427e-02 9.88858457e-01]
 [4.02315299e-03 9.95976847e-01]
 [9.29672167e-01 7.03278326e-02]
 [1.15218988e-07 9.99999885e-01]
 [9.27547364e-06 9.99990725e-01]
 [2.15755050e-03 9.97842449e-01]
 [9.99999914e-01 8.64077000e-08]
 [6.61097355e-01 3.38902645e-01]
 [2.12310345e-05 9.99978769e-01]
 [6.47489924e-03 9.93525101e-01]
 [4.85692559e-04 9.99514307e-01]
 [6.16822679e-04 9.99383177e-01]
 [2.35213027e-01 7.64786973e-01]
 [5.76591859e-03 9.94234081e-01]
 [9.54840595e-01 4.51594055e-02]
 [2.26087479e-02 9.77391252e-01]
 [1.23077594e-04 9.99876922e-01]
 [2.73490505e-01 7.26509495e-01]
 [9.99993257e-01 6.74292016e-06]
 [2.48537673e-01 7.51462327e-01]
 [2.79637771e-04 9.99720362e-01]
 [3.41778926e-05 9.99965822e-01]
 [5.45691550e-02 9.45430845e-01]
 [2.48209019e-01 7.51790981e-01]
 [1.10846706e-04 9.99889153e-01]
 [2.22481709e-05 9.99977752e-01]
 [9.95962325e-01 4.03767533e-03]
 [9.99741830e-01 2.58169951e-04]
 [5.48318990e-01 4.51681010e-01]
 [1.55247779e-02 9.84475222e-01]
 [9.99999626e-01 3.74347853e-07]
 [1.28909148e-02 9.87109085e-01]
 [4.49259242e-05 9.99955074e-01]
 [2.11867470e-05 9.99978813e-01]
 [9.66417989e-01 3.35820110e-02]
 [4.77700734e-03 9.95222993e-01]
 [4.22329894e-01 5.77670106e-01]]