func TestKMeansOnNumImagesData(t *testing.T) {
numClusters := 10
lines, err := utils.ReadLines("../demo/data/MNISTnumImages5000.txt")
if err != nil {
panic(err)
}
data := utils.StringArrayToFloatArray(lines)
start := time.Now()
clusterer := clusterer.NewKMeansSimple(numClusters, data)
clusterer.Run()
result := clusterer.GetCentroids()
time := time.Since(start)
totalSqDist := float64(0)
for _, vector := range data {
_, dist := utils.FindNearestDistance(vector, result)
totalSqDist += dist * dist
}
t.Log("Total Square Distance: ", totalSqDist)
t.Log("Average Square Distance: ", totalSqDist/float64(len(data)))
t.Log("Runtime(seconds): ", time.Seconds())
if len(result) != numClusters {
t.Errorf("RPHash Stream did not present the correct number of clusters.")
}
}
func TestSimpleLeastDistanceVsKmeans(t *testing.T) {
//Create fake data
var numClusters = 5
var numRows = 400
var dimensionality = 1000
data := make([][]float64, numRows, numRows)
for i := 0; i < numRows; i++ {
row := make([]float64, dimensionality, dimensionality)
for j := 0; j < dimensionality; j++ {
row[j] = rand.Float64()
}
data[i] = row
}
start := time.Now()
//Test RPHash with Fake Object
RPHashObject := reader.NewSimpleArray(data, numClusters)
simpleObject := simple.NewSimple(RPHashObject)
simpleObject.Run()
if len(RPHashObject.GetCentroids()) != numClusters {
t.Errorf("Requested %v centriods. But RPHashSimple returned %v.", numClusters, len(RPHashObject.GetCentroids()))
}
rpHashResult := RPHashObject.GetCentroids()
fmt.Println("RPHash: ", time.Since(start))
//Find clusters using KMeans
start = time.Now()
clusterer := clusterer.NewKMeansSimple(numClusters, data)
clusterer.Run()
kMeansResult := clusterer.GetCentroids()
fmt.Println("kMeans: ", time.Since(start))
var kMeansAssignment = 0
var rpHashAssignment = 0
var matchingAssignmentCount = 0
var kMeansTotalDist = float64(0)
var rpHashTotalDist = float64(0)
for _, vector := range data {
rpHashAssignment, _ = utils.FindNearestDistance(vector, rpHashResult)
kMeansAssignment, _ = utils.FindNearestDistance(vector, kMeansResult)
kMeansTotalDist += utils.Distance(vector, kMeansResult[kMeansAssignment])
rpHashTotalDist += utils.Distance(vector, rpHashResult[rpHashAssignment])
//t.Log(rpHashAssignments[i], kMeansAssignments[i]);
if rpHashAssignment == kMeansAssignment {
matchingAssignmentCount += 1
}
}
t.Log("RPHash:", rpHashTotalDist)
t.Log("KMeans:", kMeansTotalDist)
t.Log("Ratio: ", kMeansTotalDist/rpHashTotalDist)
}
func BenchmarkKMeans(b *testing.B) {
var numClusters = 5
var numRows = 4000
var dimensionality = 1000
data := make([][]float64, numRows, numRows)
for i := 0; i < numRows; i++ {
row := make([]float64, dimensionality, dimensionality)
for j := 0; j < dimensionality; j++ {
row[j] = rand.Float64()
}
data[i] = row
}
for i := 0; i < b.N; i++ {
clusterer := clusterer.NewKMeansSimple(numClusters, data)
clusterer.Run()
clusterer.GetCentroids()
}
}
func TestClustererUniformVectors(t *testing.T) {
//initilize data
var numClusters = 2
var numDataPoints = 8
var dimensionality = 4
data := make([][]float64, numDataPoints)
for i := 0; i < numDataPoints; i++ {
data[i] = make([]float64, dimensionality)
for j := 0; j < dimensionality; j++ {
data[i][j] = float64(i)
}
}
//run test
clusterer := clusterer.NewKMeansSimple(numClusters, data)
clusterer.Run()
var result = clusterer.GetCentroids()
//Test Results
if len(result) != numClusters {
t.Errorf("Clusterer created %v clusters. When %v was input for k.", len(result), numClusters)
}
if len(result[0]) != dimensionality {
t.Errorf("Cluster dimensionalioty of %v does not match the dimensionality of the input data, %v.", len(result[0]), dimensionality)
}
expectedResults := make([]float64, numClusters)
expectedResults[0] = 1.5 // (0+1+2+3)/4 = 1.5
expectedResults[1] = 5.5 // (4+5+6+7)/4 = 5.5
for i := 0; i < numClusters; i++ {
for j := 0; j < dimensionality; j++ {
if result[i][j] != expectedResults[i] {
t.Errorf("Data did not cluster as expected. Data: %v, Clusters: %v. Failure at %v, %v.", data, result, i, j)
}
}
}
}