Learning the parameters of the prediction function on one data set and testing on the same data set is a wrong method: because a model only repeats the labels of the samples that have just been trained, the result is very high in this case, but it is impossible to predict when encountering samples that have not been trained.This is called overfitting.In order to avoid over-fitting, a common practice is to keep some samples as test sets (X_test, y_test) when conducting a (supervised) machine learning experiment.
chinese document: http://sklearn.Apache cn.org/cn/stable/modules/cross _ validation.html
english document: http://sklearn.Apache cn.org/en/stable/modules/cross _ validation.html
Official Document: http://scikit-learn.org/stable/
github: https://github.com/apachecn/scikit-learn-doc-zhStar, we have been working hard)
Contributor: https://github.com/apachecn/scikit-learn-doc-zh# Contributor
About Us: http://www.apachecn.org/organization/209.html
Note: This document is being translated...
Abstract: Background In development, we often use Executors class to create thread pools to perform a large number of tasks, and use the concurrency characteristics of thread pools to improve system throughput.However, improper use of thread pool will also deplete server resources and lead to abnormal situations, such as too many blocked queue tasks in fixed thread pool and too many threads created in cache thread pool, resulting in problems
Data Source from Zhihu-"Whether it is effective is still unknown.
First Book, Getting Started
《Head first HTML&CSS》The best introductory book.After reading it twice, I have a general impression of HTML & CSS.At this time, the w3cschool is collected as a reference manual.
What is a Thread Pool In multi-threaded development, due to the large number of threads and the completion of each thread's execution for a period of time, it is necessary to create threads frequently. However, such frequent thread creation will greatly reduce the efficiency of the system, because it takes time to create threads frequently and destroy threads.In this case, people want a thread that does not need to be
Summary of Form Usage in HTML
1. Standard Form Format:<form method="" action="" enctype="multipart/form-data" > <input type="text" name="username" /> <input type="password" name="pwd" /> </form > where the value of method can make get or post different is that when get requests, the data size cannot exceed 2k, and the requested content will appear on the address bar.When post requests, the request data cannot exceed 8M but can be increased by setting.
First of all, let's think about why deep learning works well.I think the biggest reason is that it can easily implement a very high complexity model, while traditionally it is not very simple to achieve high complexity.Traditional methods:
Feature Transformation (in fact, the means are limited, because there are few kinds of practical stone functions) Non-linear model, some models represented by decision tree, adaboost, gbdt, rf, etc. These models always have some limitations if they want to increase the complexity, while deep learning does not have this problem.
1. Advantages of this method: can acquire various states of threads in the thread pool in real time thread pool size can be dynamically adjusted 2. Brief Introduction of Working Principle of Thread Pool; If the number of threads in the current thread pool is less than CoreBoost, a thread will be created to execute each task. if the number of threads in the current thread pool > = coreboost,
For example, given two values of 5 and 10, find the maximum value directly without comparison.
When it comes to finding the maximum value without comparison, you don't have to think about it. Generally, it can only be realized through bit operation.
max = a - ((a-b)&((a-b)>>31))
max = ((a+b)+|a-b|)/2
If we find the minimum value, we only need to add the two values and subtract max.