有48个预测值和对应实际值采用百位数法中的中位数进行区间预测的MATLAB代码置信区间为90%和95%评价指标为覆盖度和平均宽度

以下是一个基于MATLAB的示例代码，用于计算给定48个预测值和对应实际值的区间预测。

% 设置预测值和实际值
predictions = [101, 92, 87, 105, 98, 99, 100, 103, 94, 106, 97, 86, 107, 96, 90, 102, 104, 93, 95, 85, 108, 89, 91, 109, 88, 84, 110, 83, 82, 111, 81, 80, 112, 79, 113, 78, 77, 114, 76, 75, 115, 74, 116, 73, 117, 72];
actual_values = [100, 90, 85, 95, 100.5, 98.5, 101.2, 102.3, 92.7, 109.8, 96.5, 86.5, 108.2, 95.8, 89.5, 100.7, 103.4, 94.2, 97.1, 85.9, 107.6, 89.3, 91.4, 111.1, 87.8, 84.6, 109.9, 83.4, 82.1, 112.5, 81.3, 80.2, 110.3, 79.7, 113.9, ...
    78.2,77.5,114.6 ,76.1 ,75.8, 115.2, 74.3, 116.8, 73.1, 117.5, 72.7];

% 计算预测值和实际值的差异
differences = abs(predictions - actual_values);

% 计算中位数
median_difference = median(differences);

% 计算百分位数
percentile_90 = prctile(differences, 90);
percentile_95 = prctile(differences, 95);

% 计算覆盖度
coverage_90 = sum(differences <= percentile_90) / numel(differences);
coverage_95 = sum(differences <= percentile_95) / numel(differences);

% 计算平均宽度
average_width_90 = 2 * percentile_90;
average_width_95 = 2 * percentile_95;

% 显示结果
fprintf('90%% 置信区间：中位数 = %.2f, 覆盖度 = %.2f%%, 平均宽度 = %.2f\n', median_difference, coverage_90*100, average_width_90);
fprintf('95%% 置信区间：中位数 = %.2f, 覆盖度 = %.2f%%, 平均宽度 = %.2f\n', median_difference, coverage_95*100, average_width_95);

请确保将预测值和实际值替换为您的数据，并根据您的需求进行相应的修改。这段代码将计算90%和95%置信区间的覆盖度和平均宽度，并输出结果。