public static void tools()
{
Dictionary <string, string> tools = new Dictionary <string, string>();
tools.Add("添加因素点", "ga.addspotui()");
tools.Add("删除因素点", "ga.delspotui()");
tools.Add("确定样本点数", "ppsp.nsamples()");
tools.Add("运行分析", "ppsp.run()");
tools.Add("重置分析", "ga.resetgrid()");
tools.Add("重置网格", "ga.readygrid()");
tools.Add("关闭网格", "ga.hidegrid()");
PyCmd.toolbox("工具箱", tools);
GridAnalysis.init();
}
public static void pickpoints()
{
UserInput.pause("即将进行分等定级。");
int levelCount = Gui.InputBox("级别数量", "5").TryParseToInt32();
var values = GridAnalysis.valuecache.Values.OrderBy(x => x).ToList();
List <List <double> > grades = null;
List <double> borderValues = null;
if (UserInput.option("选择分等定级算法。", "值域均分法", "动态聚类法") == 0)
{
double max = values.Max();
double min = values.Min();
double delta = (max - min) / levelCount;
borderValues = Enumerable.Range(0, levelCount).Select(i => min + i * delta).ToList();
borderValues.Add(max + 0.001 * delta);
grades = borderValues.PairwiseSelect((x, y) => values.Where(z => x <= z && z < y).ToList()).ToList();
GradeTheme gt = new GradeTheme("value", borderValues);
MapControl.Current.GridLayer.ApplyColorTheme(gt);
}
else
{
var steps = Enumerable.Range(0, 5).ToList();
steps.Add(1000);
foreach (var step in steps)
{
var result = Dreambuild.Mathematics.DynamicGrading.Perform(values, levelCount, step);
grades = result.Item1;
borderValues = result.Item2;
GradeTheme gt = new GradeTheme("value", borderValues);
MapControl.Current.GridLayer.ApplyColorTheme(gt);
UserInput.pause(string.Format("已进行第 {0} 步聚类。", result.Item3));
}
}
UserInput.pause("即将确定样本容量。");
int totalSampleCount = Gui.InputBox("总样本点数", "100").TryParseToInt32();
var Es = grades.Select(x => x.StandardDeviation()).ToList();
var Hs = grades.Select(x => x.Sum()).ToList();
var EHs = Es.Zip(Hs, (e, h) => e * h).ToList();
double EHSum = EHs.Sum();
var proportions = EHs.Select(x => x / EHSum).ToList();
var sampleCounts = proportions.Select(x => (int)Math.Ceiling(x * totalSampleCount)).ToArray();
var symbolValues = borderValues.PairwiseSelect((x, y) => (x + y) / 2).ToArray();
UserInput.pause(string.Format("各级别样本点数：{0}\n各级别标志分值：{1}", string.Join(", ", sampleCounts), string.Join(", ", symbolValues.Select(x => x.ToString("0.00")).ToArray())));
UserInput.pause("即将开始选点。");
if (UserInput.option("选择选点算法。", "多重因素法", "单一因素法") == 0)
{
Random rand = new Random(0);
Enumerable.Range(0, levelCount).ForEach(i =>
{
var cells = GridAnalysis.valuecache.Where(z => borderValues[i] <= z.Value && z.Value < borderValues[i + 1]).Select(z => z.Key).ToList();
var searches = Enumerable.Range(0, 1000).Select(x =>
{
var pts = Enumerable.Range(0, sampleCounts[i]).Select(j => cells[rand.Next(0, cells.Count - 1)]).ToList();
return(System.Tuple.Create(pts, GridAnalysis.evaluate(pts, symbolValues[i])));
}).ToList();
var target = searches.First(x => x.Item2 == searches.Min(y => y.Item2));
target.Item1.ForEach(p => GridAnalysis.markcell(p));
});
}
else
{
Enumerable.Range(0, levelCount).ForEach(i =>
{
var cells = GridAnalysis.valuecache.Where(z => borderValues[i] <= z.Value && z.Value < borderValues[i + 1]).Select(z => z.Key).ToList();
var targets = cells.OrderBy(p => Math.Abs(GridAnalysis.valuecache[p] - symbolValues[i])).Take(sampleCounts[i]).ToList();
targets.ForEach(p => GridAnalysis.markcell(p));
});
}
}
public static void run()
{
GridAnalysis.clearmark();
GridAnalysis.readygrid();
PPSP.pickpoints();
}
