def vote(features):
"""
Takes a list of features, returns the most common feature of the list.
"""
hist = util.histogram(features)
diagnosis = sorted(hist, key=hist.get)[-1]
return diagnosis
def vote(features):
"""
Takes a list of features, returns the most common feature of the list.
"""
hist = util.histogram(features)
diagnosis = sorted(hist, key=hist.get)[-1]
return diagnosis
def main(arg):
k = int(arg.k)
results = []
dataFilePath = arg.dataFile
testDataFilePath = arg.testDataFile
searchKey = arg.propertyToLearn
header, data = knnloader.getDataFromFilename(dataFilePath)
headerIndices = util.arrayToReverseDict(header)
_, testData = knnloader.getDataFromFilename(testDataFilePath)
confusionMatrix = np.array([0, 0, 0, 0])
for queryRow in testData:
searchIndex = headerIndices[searchKey]
result = knn(
k,
searchIndex,
queryRow,
header,
data
)
cfMatrixIndex = 1 if result == "Sick" else 0
cfMatrixIndex += 2 if queryRow[searchIndex] == "Healthy" else 0
confusionMatrix[cfMatrixIndex] += 1
results.append(result)
knnloader.writeResultsToDisk(header, results, testData, "diagnoses.csv")
computedHistogram = util.histogram(results)
confusionMatrix = confusionMatrix.astype(float)
errorRate = (
confusionMatrix[2] +
confusionMatrix[1]
)/(
confusionMatrix[0] +
confusionMatrix[1] +
confusionMatrix[2] +
confusionMatrix[3]
)
accuracyRate = 1 - errorRate
sensitivity = confusionMatrix[2]/(confusionMatrix[2]+confusionMatrix[3])
specificity = confusionMatrix[1]/(confusionMatrix[1]+confusionMatrix[0])
precision = confusionMatrix[2]/(confusionMatrix[2]+confusionMatrix[0])
print "Accuracy Rate: ", accuracyRate
print "Error Rate:", errorRate
print "Sensitivity: ", sensitivity
print "Specificity: ", specificity
print "Precision: ", precision
print confusionMatrix.reshape(2, 2)
print computedHistogram
def main(arg):
k = int(arg.k)
results = []
dataFilePath = arg.dataFile
testDataFilePath = arg.testDataFile
searchKey = arg.propertyToLearn
header, data = knnloader.getDataFromFilename(dataFilePath)
headerIndices = util.arrayToReverseDict(header)
_, testData = knnloader.getDataFromFilename(testDataFilePath)
confusionMatrix = np.array([0, 0, 0, 0])
for queryRow in testData:
searchIndex = headerIndices[searchKey]
result = knn(k, searchIndex, queryRow, header, data)
cfMatrixIndex = 1 if result == "Sick" else 0
cfMatrixIndex += 2 if queryRow[searchIndex] == "Healthy" else 0
confusionMatrix[cfMatrixIndex] += 1
results.append(result)
knnloader.writeResultsToDisk(header, results, testData, "diagnoses.csv")
computedHistogram = util.histogram(results)
confusionMatrix = confusionMatrix.astype(float)
errorRate = (confusionMatrix[2] + confusionMatrix[1]) / (
confusionMatrix[0] + confusionMatrix[1] + confusionMatrix[2] +
confusionMatrix[3])
accuracyRate = 1 - errorRate
sensitivity = confusionMatrix[2] / (confusionMatrix[2] +
confusionMatrix[3])
specificity = confusionMatrix[1] / (confusionMatrix[1] +
confusionMatrix[0])
precision = confusionMatrix[2] / (confusionMatrix[2] + confusionMatrix[0])
print "Accuracy Rate: ", accuracyRate
print "Error Rate:", errorRate
print "Sensitivity: ", sensitivity
print "Specificity: ", specificity
print "Precision: ", precision
print confusionMatrix.reshape(2, 2)
print computedHistogram
