利用西瓜数据集进行实验，训练生成决策树并进行测试，将数据集的1到7与10到17行作为训练集，剩下的为测试集。1)实现决策树学习算法，要求度量样本集合纯度指标采用“信息增益”，画出决策树并给出相应的测试结果。给出matlab代码

好的，以下是利用Matlab实现决策树学习算法并进行测试的代码：

% 载入西瓜数据集
data = xlsread('watermelon.xlsx');
train_data = [data(1:7,:);data(10:17,:)]; % 训练集
test_data = data(8:9,:); % 测试集

% 定义信息熵计算函数
function [Ent] = entropy(data)
    label = unique(data(:,end));
    n = size(data,1);
    Ent = 0;
    for i = 1:length(label)
        p = sum(data(:,end) == label(i))/n;
        Ent = Ent - p*log2(p);
    end
end

% 定义信息增益计算函数
function [Gain] = info_gain(data, feature)
    n = size(data,1);
    Ent_D = entropy(data);
    value = unique(data(:,feature));
    Gain = Ent_D;
    for i = 1:length(value)
        Dv = data(data(:,feature)==value(i),:);
        Ent_Dv = entropy(Dv);
        p = size(Dv,1)/n;
        Gain = Gain - p*Ent_Dv;
    end
end

% 定义选择最优特征函数
function [best_feature, best_gain] = choose_feature(data)
    n_feature = size(data,2)-1;
    best_gain = 0;
    for i = 1:n_feature
        Gain = info_gain(data, i);
        if Gain > best_gain
            best_feature = i;
            best_gain = Gain;
        end
    end
end

% 定义构建决策树函数
function [Tree] = create_tree(data)
    label = unique(data(:,end));
    if length(label) == 1 % 判断是否为叶节点
        Tree.label = label;
        Tree.feature = [];
        Tree.child = [];
        return;
    end
    [best_feature, best_gain] = choose_feature(data); % 选择最优特征
    Tree.feature = best_feature;
    Tree.child = {};
    value = unique(data(:,best_feature));
    for i = 1:length(value)
        Dv = data(data(:,best_feature)==value(i),:);
        if isempty(Dv) % 处理空集
            sub_tree = struct('label', mode(data(:,end)), 'feature', [], 'child', []);
        else
            sub_tree = create_tree(Dv);
        end
        Tree.child{i} = sub_tree;
    end
    Tree.label = [];
end

% 构建决策树
Tree = create_tree(train_data);

% 定义决策函数
function [label] = decision_tree(data, Tree)
    if ~isempty(Tree.label) % 判断是否为叶节点
        label = Tree.label;
        return;
    end
    feature = Tree.feature;
    value = data(feature);
    child = Tree.child;
    for i = 1:length(child)
        if value == i
            label = decision_tree(data, child{i});
            return;
        end
    end
end

% 进行测试
n = size(test_data,1);
accuracy = 0;
for i = 1:n
    label = decision_tree(test_data(i,:), Tree);
    if label == test_data(i,end)
        accuracy = accuracy + 1;
    end
end
accuracy = accuracy/n;
disp(['测试集准确率为：', num2str(accuracy)]);

该代码实现了决策树学习算法，并采用信息增益作为样本集合纯度指标。同时，该代码还实现了决策树的构建和测试，最终输出测试集准确率。