speech noise reduction

频繁项集挖掘算法，能在大量局部特征中发现频繁空间配置，这些空间配置可以作为词语，加入到特征包中进行分类，实现图像分类。

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.)

1基于MapReduce的K-Means聚类算法， 2基于MapReduce的分类算法 3MapReduce的频繁项集挖掘算法(1 K-Means clustering algorithm based on MapReduce, 2 classification algorithm based on MapReduce 3 Mining Algorithm for frequent itemsets based on MapReduce)

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. )

利用一维卷积神经网络将叶子进行分类，里面包含的有数据(Classification of leaves using one dimensional convolution neural network)

CHKS光滑孪生支持向量机程序, 采用CHKS光滑函数逼近无约束孪生支持向量机的不可微部分，得到一类光滑的孪生支持向量机。(CHKS smooth twin support vector machine program)

说明：  内附源数据、代码及word。代码包括：平稳性检验、协整检验、滞后阶数的确定、VAR 模型的拟合、脉冲响应分析、VAR 模型的预测(Stationarity test, co integration test, determination of lag order, VAR model fitting, impulse response analysis, VAR model prediction)

数据挖掘算法，高效用项集挖掘算法，加权频繁项集挖掘，HUIMINER算法，源码中有详细注释

时间序列法，通过过去数据来建立相应模型来预测未来数据(Time series, using past data to establish corresponding models to predict future data)

主要利用Python软件，利用KNN算法对垃圾邮件进行分类(This paper mainly uses Python software to classify spam mail by using KNN algorithm)