二 使用时为matlab+spider+Weka；因为spider中的一些算法引用了Weka，比如j48

安装注意：

1 matlab7（R14）

  6.5版本对java的支持不够，还没有开发javaclasspath等函数

??? Undefined function or variable 'javaclasspath'.
??? Undefined function or variable 'javaaddclasspath'.

2 jre1.4.2

matlab7自带的是1.4.2；matlab6自带的是1.3.可以在D:\MATLAB7\sys\java\jre\win32下看到。如果装了matlab7，使用它自带的1.4.2就可以了，尤其不要使用1.6，因为1.6太新了，matlab还不支持。可以在Matlab下使用 version -java查看JVM版本。

如果你想使用1.5的话，C:\Program Files\Java\jre1.5.0_10；把jre1.5.0_10这个文件夹拷贝到D:\MATLAB7\sys\java\jre\win32下，然后增加环境变量MATLAB_JAVA：D:\MATLAB7\sys\java\jre\win32\jre1.5.0_10。这一步如果有问题的话，重启Matlab会给出错误提示。找不到什么什么文件...

3 Weka3.4.10

  使用weka版本低一些即可，高的不行，因为高版本的weka可能是用高版本的jvm支持的。

我使用的组合是 matlab7（R14）+jre1.4.2（matlab7自带的，不需要任何设置）+Weka3.4.10

三 使用方法

1 下载spider，有core和extra两个压缩包，把他们解压到同一个文件夹spider下面，然后放到$matlabroot\toolbox下面

2下载weka3.4.10，找到weka.jar放到$matlabroot\java\jar下面

3 启动Matlab打开$matlabroot\toolbox\spider\use_spider.m运行

提示spider的一些信息和 WEKA support enabled!表示成功了。

然后可以使用 help spider命令查看信息，他的功能列出如附录，然后就可以训练了。

四 一个简单的例子

X=rand(50)-0.5; Y=sign(sum(X,2));
dtrain=data(X,Y);
%生成训练集，也可以使用load()从文件读取

model=train(svm,dtrain));
%使用函数train（）训练模型

rtest=test(dtest,model);
%使用训练好的模型对验证集dtest测试，返回测试结果

五 附录spider信息

最新spider Version 1.71 (24/7/2006)

Basic library objects.
    data         - Storing input data and output results
    data_global  - Implementation of data object that limits memory overhead
    algorithm    - Generic algorithm object
    group        - Groups sets of objects together (algorithms or data)
    loss         - Evaluates loss functions
    get_mean     - Takes mean loss over groups of algs
    chain        - Builds chains of objects: output of one to input of another
    param        - To train and test different hyperparameters of an object
    cv           - Cross validation using objects given data
    kernel       - Evaluates and caches kernel functions
    distance     - Evaluates and caches distance functions

  Statistical Tests objects.
    wilcoxon     - Wilcoxon test of statistical significance of results
    corrt_test   - Corrected resampled t-test - for dependent trials

  Dataset objects.
    spiral       - Spiral dataset generator.
    toy          - Generator of dataset with only a few relevant features
    toy2d        - Simple 2d Gaussian problem generator
    toyreg       - Linear Regression with o outputs and n inputs

  Pre-Processing objects
    normalize    - Simple normalization of data
    map          - General user specified mapping function of data

  Density Estimation objects.
    parzen       - Parzen's windows kernel density estimator
    indep        - Density estimator which assumes feature independence
    bayes        - Classifer based on density estimation for each class
    gauss        - Normal distribution density estimator
                     
  Pattern Recognition objects.
    svm          - Support Vector Machine (svm)
    c45          - C4.5 for binary or multi-class
    knn          - k-nearest neighbours
    platt        - Conditional Probability estimation for margin classifiers
    mksvm        - Multi-Kernel LP-SVM
    anorm        - Minimize the a-norm in alpha space using kernels
    lgcz         - Local and Global Consistent Learner
    bagging      - Bagging Classifier
    adaboost     - ADABoost method
    hmm          - Hidden Markov Model
    loom         - Leave One Out Machine
    l1           - Minimize l1 norm of w for a linear separator
    kde          - Kernel Dependency Estimation: general input/output machine
    dualperceptron       - Kernel Perceptron
    ord_reg_perceptron   - Ordinal Regression Perceptron (Shen et al.)
    splitting_perceptron - Splitting Perceptron (Shen et al.)
    budget_perceptron    - Sparse, online Pereceptron (Crammer et al.)
    randomforest - Random Forest Decision Trees         WEKA-Required
    j48          - J48 Decision Trees for binary        WEKA-Required

  Multi-Class and Multi-label objects.
    one_vs_rest  - Voting method of one against the rest (also for multi-label)
    one_vs_one   - Voting method of one against one
    mc_svm       - Multi-class Support Vector Machine by J.Weston
    c45          - C4.5 for binary or multi-class
    knn          - k-nearest neighbours
                
  Feature Selection objects.
    feat_sel     - Generic object for feature selection + classifier
    r2w2_sel     - SVM Bound-based feature selection
    rfe          - Recursive Feature Elimination (also for the non-linear case)
    l0           - Dual zero-norm minimization (Weston, Elisseeff)
    fsv          - Primal zero-norm based feature selection (Mangasarian)
    fisher       - Fisher criterion feature selection
    mars         - selection algorithm of Friedman (greedy selection)
    clustub      - Multi-class feature selection using spectral clustering
    mutinf       - Mutual Information for feature selection.
       
  Regression objects.
    svr          - Support Vector Regression
    gproc        - Gaussian Process Regression
    relvm_r      - Relevance vector machine
    multi_rr     - (possibly multi-dimensional) ridge regression   
    mrs          - Multivariate Regression via Stiefel Constraints      
    knn          - k-nearest neighbours
    multi_reg    - meta method for independent multiple output regression
    kmp          - kernel matching pursuit
    kpls         - kernel partial least squares
    lms          - least mean squared regression [now obselete due to multi_rr]
    rbfnet       - Radial Basis Function Network (with moving centers)
    reptree      - Reduced Error Pruning Tree       WEKA-Required

  Model Selection objects.
    gridsel      - select parameters from a grid of values
    r2w2_sel     - Selecting SVM parameters by generalization bound
    bayessel     - Bayessian parameter selection

  Unsupervised objects.
    one_class_svm - One class SVM
    kmeans       - K means clustering
    kvq          - Kernel Vector Quantization
    kpca         - Kernel Principal Components Analysis
    ppca         - Probabilistic Principal Component Analysis
    nmf          - Non-negative Matrix factorization
    spectral     - Spectral clustering
    mrank        - Manifold ranking
    ppca         - Probabilistic PCA

  Reduced Set and Pre-Image objects.
    pmg_mds      - Calculate Pre-Images based on multi-dimensional scaling
    pmg_rr       - Calculate Pre-Images based on learning and ridge regression
    rsc_burges   - Bottom Up Reduced Set; calculates reduced set based on gradient descent
    rsc_fp       - Bottom Up Reduced Set; calculates reduced set for rbf with fixed-point iteration schemes
    rsc_mds      - Top Down Reduced Set; calculates reduced set with multi-dimensional scaling
    rsc_learn    - Top Down Reduced Set; calculates reduced set with ridge regression
    rss_l1       - Reduced Set Selection via L1 penalization
    rss_l0       - Reduced Set Selection via L0 penalization
    rss_mp       - Reduced Set Selection via matching pursuit