logistic regression on the SVM’s scores, fit by an additional cross-validation on the training data. In the multiclass case, this is extended as per

PROBABILITY AND STATISTICAL INFERENCE，PROBABILITY AND STATISTICAL INFERENCE，概率与统计推断

Lawley-Donovan-素数概率 用法：./{编译程序} [N] 复杂度] 其中 N 是要停止的最大整数，复杂度是停止计算的因子（2，对于名义方法。） ============================== 来自 Lawley 论文的摘要：Lawley 素性检验是一个常数时间的概率素性判断一个数是合数还是可能素数的测试算法。 让 n 是要测试的正整数。 如果 n 是偶数，则该数字是合数。 如果 n 是奇数，则该数“可能”是素数； 掷 d6，如果得到 1 或 4*，则返回素数。 否则，返回复合。 该算法取得了惊人的实验结果：对于前 100,000 个正整数，它实现了低至 6% 的误报率和低至 13% 的误报率。 这些数字使其非常接近领先的概率素性测试算法（其中大多数实现了 0% 的全面错误率），而复杂性只是其中的一小部分。 *由现在著名的“多诺万修订版”修改为 1。

This book is written for high school and college students learning about probability for the first time. It will appeal to the reader who has a healthy level of enthusiasm for understanding how and why the various results of probability come about.All of the standard introductory topics in probability are covered: combinatorics, the rules of probability, Bayes’ theorem, expectation value, variance, probability density, common distributions, the law of large numbers, the central limit theorem, correlation, and regression. Calculus is not a prerequisite, although a few of the problems do involve calculus. These are marked clearly. The book features 150 worked-out problems in the form of examples in the text and solved problems at the end of each chapter. These problems, along with the discussions in the text, will be a valuable resource in any introductory probability course, either as the main text or as a helpful supplement.

Breiman的probability是目前我见过的最好的统计理论教材。通俗易懂，涵盖的也比较全面。这版质量非常好。

经典教材 Sheldon M. Ross This updated text provides a superior introduction to applied probability and statistics for engineering or science majors. Ross emphasizes the manner in which probability yields insight into statistical problems; ultimately resulting in an intuitive understanding of the statistical procedures most often used by practicing engineers and scientists. Real data sets are incorporated in a wide variety of exercises and examples throughout the book, and this emphasis on data motivates the probability coverage.

具有TensorFlow的贝叶斯神经网络 如我的论文所述，该存储库考虑了贝叶斯人工神经网络的实现。 实施的中心是模块 ，其中我们的许多技术工作都受到启发。 该存储库的主要结构如下：

Probability for Machine Learning Discover How To Harness Uncertainty With Python by Jason Brownlee 28 step-by-step lessons, 312 pages. distributions, estimation, more. 概率是机器学习的基石。没有机器学习，你就无法对机器学习产生深刻的理解和应用。 切开方程、希腊字母和混淆，并发现您需要了解的概率主题。 使用清晰的解释、标准的 Python 库和分步教程课程，您将发现概率对机器学习、贝叶斯概率、熵、密度估计、最大可能性等的重要性。

国外一位叫Feller的大神出的概率论教材，适合需要对概率论知识深入了解的人使用。

Probability: Theory and Examples Rick Durrett, Duke U.