// Train converts the Attributes into equivalently named FloatAttributes,
// leaves FloatAttributes unmodified and processes
// CategoricalAttributes as follows.
//
// If the CategoricalAttribute has two values, one of them is
// designated 0.0 and the other 1.0, and a single identically-named
// FloatAttribute is returned.
//
// If the CategoricalAttribute has more than two (n) values, the Filter
// generates n FloatAttributes and sets each of them if the value's observed.
func (f *FloatConvertFilter) Train() error {
for _, a := range f.attrs {
if ac, ok := a.(*base.CategoricalAttribute); ok {
vals := ac.GetValues()
if len(vals) <= 2 {
nAttr := base.NewFloatAttribute(ac.GetName())
fAttr := base.FilteredAttribute{ac, nAttr}
f.converted = append(f.converted, fAttr)
f.twoValuedCategoricalAttributes[a] = true
} else {
if _, ok := f.nValuedCategoricalAttributeMap[a]; !ok {
f.nValuedCategoricalAttributeMap[a] = make(map[uint64]base.Attribute)
}
for i := uint64(0); i < uint64(len(vals)); i++ {
v := vals[i]
newName := fmt.Sprintf("%s_%s", ac.GetName(), v)
newAttr := base.NewFloatAttribute(newName)
fAttr := base.FilteredAttribute{ac, newAttr}
f.converted = append(f.converted, fAttr)
f.nValuedCategoricalAttributeMap[a][i] = newAttr
}
}
} else if ab, ok := a.(*base.FloatAttribute); ok {
fAttr := base.FilteredAttribute{ab, ab}
f.converted = append(f.converted, fAttr)
} else if af, ok := a.(*base.BinaryAttribute); ok {
newAttr := base.NewFloatAttribute(af.GetName())
fAttr := base.FilteredAttribute{af, newAttr}
f.converted = append(f.converted, fAttr)
} else {
return fmt.Errorf("Unsupported Attribute type: %v", a)
}
}
return nil
}
func main() {
// Instances can be read using ParseCsvToInstances
rawData, err := base.ParseCSVToInstances("../datasets/iris_headers.csv", true)
if err != nil {
panic(err)
}
// Instances can be printed, and you'll see a human-readable summary
// if you do so. The first section is a line like
// Instances with 150 row(s) and 5 attribute(s)
//
// It next prints all the attributes
// FloatAttribute(Sepal length)
// FloatAttribute(Sepal width)
// FloatAttribute(Petal length)
// FloatAttribute(Petal width)
// CategoricalAttribute([Iris-setosa Iris-versicolor Iris-viriginica])
// The final attribute has an asterisk (*) printed before it,
// meaning that it is the class variable. It then prints out up to
// 30 rows which correspond to those attributes.
// 5.10 3.50 1.40 0.20 Iris-setosa
// 4.90 3.00 1.40 0.20 Iris-setosa
fmt.Println(rawData)
// If two decimal places isn't enough, you can update the
// Precision field on any FloatAttribute
if attr, ok := rawData.GetAttr(0).(*base.FloatAttribute); !ok {
panic("Invalid cast")
} else {
attr.Precision = 4
}
// Now the first column has more precision
fmt.Println(rawData)
// We can update the set of Instances, although the API
// for doing so is not very sophisticated.
rawData.SetAttrStr(0, 0, "1.00")
rawData.SetAttrStr(0, rawData.ClassIndex, "Iris-unusual")
fmt.Println(rawData)
// There is a way of creating new Instances from scratch.
// Inside an Instance, everything's stored as float64
newData := make([]float64, 2)
newData[0] = 1.0
newData[1] = 0.0
// Let's create some attributes
attrs := make([]base.Attribute, 2)
attrs[0] = base.NewFloatAttribute()
attrs[0].SetName("Arbitrary Float Quantity")
attrs[1] = new(base.CategoricalAttribute)
attrs[1].SetName("Class")
// Insert a standard class
attrs[1].GetSysValFromString("A")
// Now let's create the final instances set
newInst := base.NewInstancesFromRaw(attrs, 1, newData)
fmt.Println(newInst)
}
func (m *OneVsAllModel) generateAttributes(from base.FixedDataGrid) map[base.Attribute]base.Attribute {
attrs := from.AllAttributes()
classAttrs := from.AllClassAttributes()
if len(classAttrs) != 1 {
panic("Only 1 class Attribute is supported!")
}
ret := make(map[base.Attribute]base.Attribute)
for _, a := range attrs {
ret[a] = a
for _, b := range classAttrs {
if a.Equals(b) {
cur := base.NewFloatAttribute(b.GetName())
ret[a] = cur
}
}
}
return ret
}
