lstm

说明：  深入理解Spark核心技术书籍，书籍对spark进行了深入讲解，并对spark源码进行了剖析(In-depth understanding of Spark core technology books, books on Spark in-depth explanation, and analysis of spark source code)

Dijkstra算法可以解决voronoi图中求解最短路径(Dijkstra' s algorithm can solve voronoi diagram to solve the shortest path)

说明：  深入理解Spark核心技术书籍，书籍对spark进行了深入讲解，并对spark源码进行了剖析(In-depth understanding of Spark core technology books, books on Spark in-depth explanation, and analysis of spark source code)

基于KPCA的故障诊断matlab代码。包括t^2统计量和q统计量(Fault Diagnosis Matlab Code Based on KPCA)

一个介绍最优化方法课件,很详细,从最简单的一维搜索到共轭梯度,最速下降法等等,都有说明(An introduction to the optimization method courseware, very detailed, from the simplest one-dimensional search to conjugate gradient, the steepest descent method, etc.)

按最大相关-最小冗余对特征排序,里面包括求特征间的互信息(Sort the features according to the maximum correlation minimum redundancy)

说明：  多尺度小波分析，用于分析时间多尺度周期。包括相应小波包及代码，可作全谱图，方差图，功率谱，全局谱，并附带相关置信区间(Multiscale wavelet analysis is used to analyze time multiscale period. Including the corresponding wavelet packet and code, it can make full spectrum, variance, power spectrum, global spectrum, with relevant confidence interval)

最优化方法之近端梯度下降法，适用于L1正则项的最优化。(Proximal gradient descent)

通过PYTHON编写的爬虫程序，爬取Geoq的人口密度数据(The population density data of Geoq is crawled by a crawler program written in PYTHON)

哈弗曼编码，构造哈夫曼树，构造哈夫曼树非常简单，将所有的节点放到一个队列中，用一个节点替换两个频率最低的节点，新节点的频率就是这两个节点的频率之和。这样，新节点就是两个被替换节点的父节点了。如此循环，直到队列中只剩一个节点（树根）。(Hoffman coding, Huffman tree structure, Huffman tree structure is very simple, all the nodes in a queue, replace the two lowest frequency node with a node, the new node and the frequency of frequency is the two nodes. So, the new node is two replaced node parent node. So the cycle, until the queue is only one node (Shu Gen). )