test_lstm

分类器的性能比较与调优: 使用scikit-learn 包中的tree，贝叶斯，knn，对数据进行模型训练，尽量了解其原理及运用。 使用不同分析三种分类器在实验中的性能比较，分析它们的特点。 本实验采用的数据集为house与segment。(Performance comparison and optimization of classifiers: We use tree, Bayesian and KNN in scikit-learnpackage to train the data model and try to understand its principle and application. The performances of three classifiers are compared and their characteristics are analyzed. The data set used in this experiment is house and segment.)

多重分形去趋势波动分析法，用于不同时间序列的重分形交叉相关性分析。(Multifractal detrended fluctuation analysis)

说明：  深度学习框架Keras入门资料，里面的代码包括课件和DEMO有利于新书入门学习，简单易懂(Keras Introductory Information of Deep Learning Framework, which includes courseware and DEMO, is helpful for introductory learning of new books. It is easy to understand.)

Matrix QR decomposition of digital analysis of the more difficult procedures, it is worth a look

請用星號( )排列出下列的 p箭 星號 D案,輸入 Y料 ne txt中的第一行為箭 的上下高度(必為奇數),第二行為 D案的總寬度()

说明：  用来作分类识别的KNN算法,非常好用非常好用(A very useful nearest neighbor algorithm)

说明：  聚类评价指标的各种说明，非常详细，请仔细阅读。(Cluster evaluation indicators of various descriptions, very detailed.)

常用数值算法源码，包括二分法、复化辛卜生公式、改进欧拉法、高斯-赛德尔迭代法、拉格郎日插值多项式、列主元高斯消去法、龙贝格算法、龙格-库塔算法、幂法、牛顿迭代法、牛顿值多项式、四阶阿当姆斯预测-校正公式、雅可比迭代法、自适应梯形公式（变步长）、最小二乘法 个人由于需要编写，完全可用(Common numerical algorithm source code, including dichotomy, complex Simpson formula, improved Euler method, Gauss-Seidel iteration method, Lagrange interpolation polynomial, column principal element Gauss elimination method, Runberg algorithm, Runge-Kutta algorithm, power method, Newton iteration method, Newton value polynomial, fourth-order Adams prediction-correction formula, Jacobi iteration method, adaptive ladder formula (variable) Step Length and Least Square Method Individuals are fully available because they need to write)

自己写的一个分形程序,支持动态编译表达式()

三维风场模拟.利用Kaimal spectrum结合FFT进行风场模拟，生成风速时程得进行必要参数的定义。(3D Turbulent Wind Generation。 Generation of three-dimensional turbulent wind fields, by employing a Kaimal spectrum and IEC-based coherence function. )