QA Official

【Scikit-Learn Chinese Document] Cross Validation-Model Selection and Evaluation-User Guide | ApacheCN

https://qaofficial.com/post/2019/04/21/70501-scikit-learn-chinese-document-cross-validation-model-selection-and-evaluation-user-guide-apachecn.html 2019-04-21
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...

【Scikit-Learn Chinese Document] hyperparameter of Optimization Estimator-Model Selection and Evaluation-User Guide | ApacheCN

https://qaofficial.com/post/2019/04/21/70523-scikit-learn-chinese-document-hyperparameter-of-optimization-estimator-model-selection-and-evaluation-user-guide-apachecn.html 2019-04-21
chinese document: http://sklearn.Apache cn.org/cn/stable/modules/grid _ search.html english document: http://sklearn.Apache cn.org/en/stable/modules/grid _ search.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 3.2. Adjust the hyperparameter of the estimator hyperparameter, i.e. parameters that are not directly learned in the estimator.In the scikit-learn package, they are passed as parameters of constructors in the estimator class.

Artificial Intelligence (3)- Model Evaluation and Parameter Adjustment

https://qaofficial.com/post/2019/04/21/70520-artificial-intelligence-3-model-evaluation-and-parameter-adjustment.html 2019-04-21
1.pipeline pipeline utilization The concept of pipeline can be abstracted from here: a thing that needs to be repeated is cut into different stages, and each stage is responsible by an independent unit.All objects to be executed enter the job queue in turn. The root of the application of pipeline intelligence in machine learning lies in the reuse of parameter sets on new data and on computers. 2.

Clustering-Spectral Clustering Algorithm and Python Implementation

https://qaofficial.com/post/2019/04/21/70477-clustering-spectral-clustering-algorithm-and-python-implementation.html 2019-04-21
spectral cluster can be regarded as an improved K-Kmeans clustering algorithm.It is often used for image segmentation.The disadvantage is that the number of clusters needs to be specified and it is difficult to construct a suitable similarity matrix.Its advantage is simple and easy to implement.Compared with Kmeans, it is more suitable to process high-dimensional data. core idea Construct the similarity matrix (graph) of sample points and cut the graph into

Data Mining: Synonymous Things

https://qaofficial.com/post/2019/04/21/70572-data-mining-synonymous-things.html 2019-04-21
What are synonyms? Ha ha, Yuan inside has a joke: 40 synonyms for wife1. Spouse 2, wife 3, wife 4, wife 5, wife 6, lover 7, wife 8, daughter-in-law 9, son 10, My wife11, wife 12, wife 13, children his mother 14, children his mother 15, wife 16, wife 17, distillers’ grains 18, baby 19, britney 20, honey 21, Wife 22, darling23, women 24, horse 25, housewife 26, hostess 27, finance minister 28, AN UNREASONABLE FELLOW 29, lady 30, elder sister 31, family members 32, 33 in the house, the other half 34, female head of household35.

General Overview of Remote Sensing Image Classification Based on Depth Learning

https://qaofficial.com/post/2019/04/21/70655-general-overview-of-remote-sensing-image-classification-based-on-depth-learning.html 2019-04-21
QQ for mutual communication: 1049974028 Blog Reprinted from: https://blog.csdn.net/qq_40116035/article/details/81414835There are two traditional classification methods: supervised classification and unsupervised classification.Supervised classification requires a priori knowledge of the classification of the area to be classified, i.e. the training area of all kinds of ground objects to be distinguished should be selected from the studied area for establishing the discrimination function.Commonly used supervised classification methods include: K nearest neighbor method, Mahalanobis distance classification, maximum likelihood method, etc.

In machine learning, there are three methods to obtain training set and test set from sample set.

https://qaofficial.com/post/2019/04/21/70540-in-machine-learning-there-are-three-methods-to-obtain-training-set-and-test-set-from-sample-set..html 2019-04-21
1. Why should training set and test set be separated In machine learning, we choose the learner by evaluating the Generalization error of the learner.The specific method is as follows: We need to produce learners from the training set data, and then use the test set to test the discriminative ability of the learners to new samples, and use the test error on the test set as the approximation of Generalization error to select learners.

Install vmware-tools to Solve Difficult and Miscellaneous Diseases!

https://qaofficial.com/post/2019/04/21/70624-install-vmware-tools-to-solve-difficult-and-miscellaneous-diseases.html 2019-04-21
Install the virtual machine under win, and then I used ubuntu. When installing VMware-tools, I couldn't find VMWare Tools-10.1.6-5214329. tar.gz on the desktop. After checking various tutorials on the Internet, it's not good to use them (it's not good to mount them first, then mount them). Let's use a simple method to solve this problem. The most difficult thing for anyone who has not studied linux to do things under ubuntu is that they are not familiar with the command line.

Selection Sorting and Bubble Sort Thought and Code Implementation in Java

https://qaofficial.com/post/2019/04/21/110545-selection-sorting-and-bubble-sort-thought-and-code-implementation-in-java.html 2019-04-21
Select Sort Basic idea of selecting sorting (assuming sorting from large to small): Initialize an array: int[] array={n data} Sort 1: Take out the element with index 0, and compare it with every subsequent element. If it is smaller than the element, it will not move. If it is larger than the element, it will exchange the values of the two elements, and then compare them in turn to the

Test Set Selection Method in Model Evaluation

https://qaofficial.com/post/2019/04/21/70510-test-set-selection-method-in-model-evaluation.html 2019-04-21
Preface The purpose of evaluating the trained model is to test whether the trained model has good generalization ability.Therefore, a "test set" should be used to test the learner's discrimination ability to new samples, and then the "test error" on the test set should be taken as the approximation of Generalization error.Generally, we assume that the test sample is also obtained by independent and identically distributed sampling from the real distribution of the sample.