Simple script to calculate mean, standard deviation, variance and optionally the gaussian probability for a given target value inside a given set of elements, that can be divided in boolean classes to calculate the statistics for each subset. Was designed as help for calculus in Naïve Bayes algorithm exercises

gaussian_statistics.py

# Libraries
import math
import sys

# Constants
MIN_ARGS   = 1

# Parameters
values     = None
values_yes = None
values_no  = None
target     = None
verbose    = False

# Help message
def help(exit_code=1):
	print("Usage: python",args[0],"{<values> [<target>|{<indexes_yes>}]|-e <exercise> [-t <target>]} [-v] [-h]\n")
	print("Given a list of comma-separated values, and a target, returns "
	      "the gaussian probability for the target value\n")
	print("If given indexes (comma-separated) for the yes class, it will "
	      "calculate gaussian probability for values whose class=no and "
	      "class=yes (indexes start by 1)\n")
	print("If specified -e, you can select an exercise, where values are "
	      "already defined\n""")
	print("Specify -v to print all values in all classes, -h to see this "
	      "help message")
	sys.exit(exit_code)

# Argument parsing
args = sys.argv
if len(args)-1 < MIN_ARGS:
	print("ERROR: Not enough arguments")
	help()

# Check options
if "-h" in args:
	help(0)
if "-v" in args:
	verbose = True
if args[1] == "-e":
	# Exercise mode
	# Definitions
	exercises = {
		"theory": {
			"values" : [83,70,68,64,69,75,75,72,81,85,80,65,72,71]
		},
		"problems": {
			"values" : [n*1000 for n in [125,100,70,120,60,220,75,95,85,90]]
		}
	}
	exercises["problems"]["values_yes"] = exercises["problems"]["values"][:7]
	exercises["problems"]["values_no"] = exercises["problems"]["values"][7:]
	exercises["theory"]["values_yes"] = exercises["theory"]["values"][:9]
	exercises["theory"]["values_no"] = exercises["theory"]["values"][9:]
	# Help
	def show_exercises():
		print("Available exercises are:")
		print(list(exercises.keys()))
		sys.exit(3)
	# Exercise selection
	if len(args)-1 < 2:
		print("ERROR: Not enough arguments, please specify an exercise")
		show_exercises()
	if args[2] not in exercises.keys():
		print("ERROR: Exercise",args[2],"not defined")
		show_exercises()
	# Assignation
	exercise = exercises[args[2]]
	values = exercise["values"]
	values_yes = exercise["values_yes"]
	values_no = exercise["values_no"]
	# Optional target assignation
	if len(args)-1 > 2:
		try:
			target = float(args[3])
		except Exception as e:
			print("ERROR: Unable to parse target to float",str(e))
			sys.exit(6)
else:
	# Manual mode
	# Values parsing
	try:
		values = [float(n) for n in args[1].split(',')]
	except Exception as e:
		print("ERROR: Unable to parse values to floats:",str(e))
		sys.exit(2)
	if len(values) == 1:
		print("ERROR: Specify more elements in the values list, please")
		sys.exit(7)
	# Target parsing
	if len(args)-1 > 1:
		try:
			target = float(args[2])
		except Exception as e:
			print("ERROR: Unable to parse target to float:",str(e))
			sys.exit(6)
	# Yes class parsing
	if len(args)-1 > 2:
		indexes_yes = args[3].split(",")
		try:
			indexes_yes = [int(n) for n in indexes_yes]
		except Exception as e:
			print("ERROR: Unable to parse indexes to ints:",str(e))
			sys.exit(4)
		try:
			values_yes = [values[i-1] for i in indexes_yes]
		except Exception as e:
			print("ERROR: Unable to pick indexes values:",str(e))
			sys.exit(5)
		values_no = list(set(values)-set(values_yes))
		if len(values_yes) < 2 or len(values_no) < 2:
			print("ERROR: Some class has less than 2 values:")
			print("Values(class=yes):",values_yes)
			print("Values(class=no):",values_no)
			sys.exit(7)
# Statistical functions
def mean(v):
	return sum(v)/len(v)

def variance(v):
	u = mean(v)
	return sum([(u-e)**2 for e in v])/(len(v)-1)

def sdev(v):
	return math.sqrt(variance(v))

def gaussian(t, v):
	s2 = variance(v)
	s = sdev(v)
	u = mean(v)
	divisor = math.sqrt(2*math.pi)*s
	exponent = ((t-u)**2)/(s2)
	return (math.e**(-exponent))/divisor

# Information
def print_info(values,target=None):
	if(verbose):
		print("Values: ",values)
	print("Mean:",mean(values),"| sdev:",sdev(values),"| Variance:",variance(values))
	if target != None:
		print("Gaussian Probability for",target,":",gaussian(target,values))
	print("")

# Results
print("ALL VALUES")
print("--------------------------------------------------------------")
print_info(values,target)
if values_yes !=None:
	print("VALUES (class=yes)")
	print("--------------------------------------------------------------")
	print_info(values_yes,target)
if values_no != None:
	print("VALUES (class=no)")
	print("--------------------------------------------------------------")
	print_info(values_no,target)