朴素贝叶斯
[TencentCloudSDKException] code:FailedOperation.ServiceIsolate message:service is stopped due to arrears, please recharge your account in Tencent Cloud requestId:3c28f05b-cbe0-449a-820b-851e09aa37ad
[En]
[TencentCloudSDKException] code:FailedOperation.ServiceIsolate message:service is stopped due to arrears, please recharge your account in Tencent Cloud requestId:60609537-68cc-4b55-9d1e-7e1adef2950c
[TencentCloudSDKException] code:FailedOperation.ServiceIsolate message:service is stopped due to arrears, please recharge your account in Tencent Cloud requestId:a7fd5c5c-cb08-4814-b9ee-641c9f53ac0e
[En]
[TencentCloudSDKException] code:FailedOperation.ServiceIsolate message:service is stopped due to arrears, please recharge your account in Tencent Cloud requestId:e2a51d3b-7010-43b8-8f8c-0e7f5ef53fd6
[TencentCloudSDKException] code:FailedOperation.ServiceIsolate message:service is stopped due to arrears, please recharge your account in Tencent Cloud requestId:792a84a7-206f-44d1-a0fd-0b297d65c428
[En]
[TencentCloudSDKException] code:FailedOperation.ServiceIsolate message:service is stopped due to arrears, please recharge your account in Tencent Cloud requestId:97ac6d6a-aed0-4cc7-bcfd-bbda15708140
朴素贝叶斯(naive Bayes )法是基于贝叶斯定理与特征条件独立假设的分类方法①。对于给定的训练数据集,首先基于特征条件独立假设学习输入输出的联合概率分布;然后基于此模型,对给定的输入 x,利用贝叶斯定理求出后验概率最大的输出 y。
激活函数
Relu
Sigmoid
1/(1+e^(-z))
Tanh
双曲正切
LeakyRelu
Batch Normalization
[TencentCloudSDKException] code:FailedOperation.ServiceIsolate message:service is stopped due to arrears, please recharge your account in Tencent Cloud requestId:13b33810-2e6d-45c8-886e-b34aa8a27c98
[En]
[TencentCloudSDKException] code:FailedOperation.ServiceIsolate message:service is stopped due to arrears, please recharge your account in Tencent Cloud requestId:6b8a8435-b37a-476f-8ed1-82405ab53eed
全连接中
将批量归 ⼀ 化层置于全连接层中的仿射变换和激活函数之间。设全连接层的输⼊为 u,权重参数和偏差参数分别为 W和 b,激活函数为 ϕ。设批量归⼀化的运算符为BN 。那么,使⽤批量归⼀化的全连接层的输出为 ϕ(BN (x))
卷积中
批量归 ⼀ 化发 ⽣ 在卷积计算之后、应⽤激活函数之前。如果卷积计算输出多个通道,我们需要对这些通道的输出分别做批量归⼀化,且每个通道都拥有独⽴的拉伸和偏移参数,并均为标量。
K-means
输入:样本集D,簇的数目k,最大迭代次数N;
输出:簇划分(k个簇,使平方误差最小);
算法步骤:
(1)为每个聚类选择一个初始聚类中心;
(2)将样本集按照最小距离原则分配到最邻近聚类;
(3)使用每个聚类的样本均值更新聚类中心;
(4)重复步骤(2)、(3),直到聚类中心不再发生变化;
(5)输出最终的聚类中心和k个簇划分。
迭代方法
梯度下降
[TencentCloudSDKException] code:FailedOperation.ServiceIsolate message:service is stopped due to arrears, please recharge your account in Tencent Cloud requestId:20b2d993-8ff1-42a4-b91e-4dd357640fd4
[En]
[TencentCloudSDKException] code:FailedOperation.ServiceIsolate message:service is stopped due to arrears, please recharge your account in Tencent Cloud requestId:3da59a55-ff2b-48df-8072-debcfea9cec9
[TencentCloudSDKException] code:FailedOperation.ServiceIsolate message:service is stopped due to arrears, please recharge your account in Tencent Cloud requestId:5a046509-6a15-4db1-9bfc-9cdb4a637e1c
[En]
[TencentCloudSDKException] code:FailedOperation.ServiceIsolate message:service is stopped due to arrears, please recharge your account in Tencent Cloud requestId:d74e3519-6e12-4fe2-ba1b-c4661f51ba7d
Adagrad
[TencentCloudSDKException] code:FailedOperation.ServiceIsolate message:service is stopped due to arrears, please recharge your account in Tencent Cloud requestId:4e920adb-b242-4bb4-afd0-7f2ad06e7489
[En]
[TencentCloudSDKException] code:FailedOperation.ServiceIsolate message:service is stopped due to arrears, please recharge your account in Tencent Cloud requestId:b8fb179d-148b-4a91-8652-ba42653f61e4
经典评价指标
Recall, precision
Recall = TPR, Recall表示实际为正的样本被判断为正样本的比例;
precision = TP/(TP+FP) 表示预测为正的样本中,实际的正样本的数量。
F1-score
F1 = 2 * P * R / (P+R)
Averaged balanced score
参考:https://www.jianshu.com/p/2eb432f31a6b
Original: https://blog.csdn.net/qq_36993032/article/details/123979054
Author: Lloaded%356
Title: 机器学习/深度学习准备资料(待更新)
原创文章受到原创版权保护。转载请注明出处:https://www.johngo689.com/561150/
转载文章受原作者版权保护。转载请注明原作者出处!