用matlab实现apriori算法

以下是用MATLAB实现Apriori算法的示例代码：

% 载入数据
data = [
    1, 3, 4;
    2, 3, 5;
    1, 2, 3, 5;
    2, 5
];

% 设置最小支持度和最小置信度
min_support = 0.5;
min_confidence = 0.7;

% 获取所有不同的项
items = unique(data(:));

% 计算项集的支持度
supports = zeros(size(items));
for i = 1:numel(items)
    supports(i) = sum(ismember(data, items(i)));
end
supports = supports / size(data, 1);

% 获取满足最小支持度的项集
frequent_items = items(supports >= min_support);

% 初始化频繁项集
frequent_sets = cell(1, 1);
frequent_sets{1} = frequent_items;

% 生成候选项集
k = 2;
while true
    candidate_sets = {};
    for i = 1:numel(frequent_sets{k-1})
        for j = i+1:numel(frequent_sets{k-1})
            candidate_set = union(frequent_sets{k-1}{i}, frequent_sets{k-1}{j});
            if numel(candidate_set) == k && ~ismember(candidate_set, candidate_sets)
                candidate_sets{end+1} = candidate_set;
            end
        end
    end
    if isempty(candidate_sets)
        break;
    end
    
    % 计算候选项集的支持度
    candidate_supports = zeros(size(candidate_sets));
    for i = 1:numel(candidate_sets)
        candidate_supports(i) = sum(all(ismember(data, candidate_sets{i}), 2));
    end
    candidate_supports = candidate_supports / size(data, 1);
    
    % 获取满足最小支持度的项集
    frequent_items = candidate_sets(candidate_supports >= min_support);
    if isempty(frequent_items)
        break;
    end
    
    % 更新频繁项集
    frequent_sets{k} = frequent_items;
    k = k + 1;
end

% 输出频繁项集
for k = 1:numel(frequent_sets)
    for i = 1:numel(frequent_sets{k})
        fprintf('Frequent %d-itemset: {%s}\n', k, num2str(frequent_sets{k}{i}));
    end
end

% 生成关联规则
rules = {};
for k = 2:numel(frequent_sets)
    for i = 1:numel(frequent_sets{k})
        for j = 1:numel(frequent_sets{k}{i})
            antecedent = setdiff(frequent_sets{k}{i}, frequent_sets{k}{i}(j));
            consequent = frequent_sets{k}{i}(j);
            confidence = sum(all(ismember(data, antecedent), 2) & ismember(data, consequent)) / sum(all(ismember(data, antecedent), 2));
            if confidence >= min_confidence
                rules{end+1} = struct('antecedent', antecedent, 'consequent', consequent, 'confidence', confidence);
            end
        end
    end
end

% 输出关联规则
for i = 1:numel(rules)
    fprintf('Rule: {%s} => {%d}, confidence: %.2f\n', num2str(rules{i}.antecedent), rules{i}.consequent, rules{i}.confidence);
end

这段代码实现了Apriori算法，可以通过调整 `min_support` 和 `min_confidence` 参数来设置最小支持度和最小置信度。该代码使用了一个 `struct` 数据结构来存储关联规则，其中 `antecedent` 和 `consequent` 表示规则的前件和后件，`confidence` 表示规则的置信度。