在 Python 中的引力搜索算法

说明：  BP神经网络算法,自动分析误差，参数迭代生成(BP neural network algorithm, automatic error analysis, iterative parameter generation)

python编程，高质量代码，高手必备python书籍,《编写高质量代码：改善Python程序的91个建议》迷你书(python programming, high quality scripts in python)

功率谱分析，labview自带WAV声信号功率谱分析(Power spectrum analysis, LabVIEW with WAV acoustic signal power spectrum analysis)

包含对时间序列进行预测的LSTM代码，有注释(The LSTM code that contains the prediction of the time series is annotated)

基于强化学习算法A3C实现的一个小案例。(A small case based on the A3C implementation of the reinforcement learning algorithm.)

This example demonstrates how to read a raw file in Google-App-Engine. 1. Change name of application in app.yml to the name of you application 2. Upload the files 3. Open the main page

说明：  Python 抓取 COM Port 未完成(Python Serach COM Port non Fininshed)

地理信息系统应用，面向ArcGIS的Python脚本编程(Geographic Information System Application, Python Script Programming for ArcGIS)

说明：  SLIC算法是simple linear iterative cluster的简称，该算法用来生成超像素（superpixel） 算法步骤： 已知一副图像大小M*N,可以从RGB空间转换为LAB空间，LAB颜色空间表现的颜色更全面 假如预定义参数K，K为预生成的超像素数量，即预计将M*N大小的图像(像素数目即为M*N)分隔为K个超像素块，每个超像素块范围大小包含[（M*N）/K]个像素 假设每个超像素区域长和宽都均匀分布的话，那么每个超像素块的长和宽均可定义为S，S=sqrt(M*N/K) 遍历操作，将每个像素块的中心点的坐标(x,y)及其lab的值保存起来，加入到事先定义好的集合中(Algorithmic steps: Given the size of an image M*N, it can be converted from RGB space to LAB space, and the color space of LAB is more comprehensive. If the predefined parameter K, K is the number of pre-generated super-pixels, i.e., the image of M*N size (the number of pixels is M*N) is expected to be divided into K super-pixel blocks, each of which contains [(M*N)/K] pixels. Assuming that the length and width of each super-pixel region are evenly distributed, the length and width of each super-pixel block can be defined as S, S = sqrt (M*N/K). The traversal operation saves the coordinates (x, y) and lab values of the central points of each pixel block and adds them to a pre-defined set.)

该程序利用python提供的libsvm库实现支持向量机的单分类，程序中通过生成数据进行算法的训练，并生成数据进行预测，结果表明算法对于已成点的排除效果非常好，对数据集的弹性非常棒。