这部分代码取自this tutorial,关于在Infer.NET库中使用学习者。我计划将它用于机器学习,但遗憾的是我无法获得此代码片段。请帮帮我。
/// <summary>
/// A mapping for the Bayes Point Machine classifier tutorial.
/// </summary>
public class ClassifierMapping
: IClassifierMapping<IList<Vector>, int, IList<string>, string, Vector>
{
public IEnumerable<int> GetInstances(IList<Vector> featureVectors)
{
for (int instance = 0; instance < featureVectors.Count; instance++)
{
yield return instance;
}
}
public Vector GetFeatures(int instance, IList<Vector> featureVectors)
{
return featureVectors[instance];
}
public string GetLabel(
int instance, IList<Vector> featureVectors, IList<string> labels)
{
return labels[instance];
}
public IEnumerable<string> GetClassLabels(
IList<Vector> featureVectors = null, IList<string> labels = null)
{
return new[] { "Female", "Male" };
}
}
提前谢谢!
答案 0 :(得分:2)
唯一远程复杂的位是yield return
:
public IEnumerable<int> GetInstances(IList<Vector> featureVectors)
{
for (int instance = 0; instance < featureVectors.Count; instance++)
{
yield return instance;
}
}
可以替换为:
public IEnumerable<int> GetInstances(IList<Vector> featureVectors)
{
var result = new List<int>();
for (int instance = 0; instance < featureVectors.Count; instance++)
{
result.Add(instance);
}
return result;
}
答案 1 :(得分:0)
这是创建一个实现IClassifierMapping
接口的Mapping类。
http://research.microsoft.com/en-us/um/cambridge/projects/infernet/docs/Infer.NET%20Learners%20-%20Bayes%20Point%20Machine%20classifiers%20-%20API%20-%20Mappings%20-%20Standard%20Data%20Format%20Mapping.aspx
此处还可以看到其他映射接口。
http://research.microsoft.com/en-us/um/cambridge/projects/infernet/docs/Infer.NET%20Learners%20-%20Bayes%20Point%20Machine%20classifiers%20-%20API%20-%20Mappings.aspx
使用此映射类,您现在可以创建贝叶斯点机器分类器。
关于代码的复杂性,你可以看看westons的回答。