44 lines
1.1 KiB
Python
44 lines
1.1 KiB
Python
# Lotto prediction with Genetic Algorithm and Mersenne Twister
|
|
# Go Namhyeon <gnh1201@gmail.com>
|
|
# 2021-04-04
|
|
# download excel data: https://dhlottery.co.kr/gameResult.do?method=byWin
|
|
|
|
import pandas as pd
|
|
import numpy as np
|
|
from geneticalgorithm2 import geneticalgorithm2 as ga
|
|
|
|
cols = [1, 13, 14, 15, 16, 17, 18]
|
|
df = pd.read_excel('excel.xlsx', skiprows=2, usecols=cols, names=[0, 1, 2, 3, 4, 5, 6])
|
|
rows = df.values[:200]
|
|
|
|
def make_num(x, n, a, b, c, min, max):
|
|
seed = int(x*a + n*b + c)
|
|
rs = np.random.RandomState(np.random.MT19937())
|
|
rs.seed(seed)
|
|
return min + round((max - min) * rs.random())
|
|
|
|
def f(X):
|
|
score = 0
|
|
|
|
for row in rows:
|
|
x = row[0]
|
|
N = row[1:]
|
|
_N = [make_num(x, n, X[0], X[1], X[2], 1, 45) for n in range(1, 7)]
|
|
result = len(list(set(N) & set(_N)))
|
|
if result > 5:
|
|
score += 100
|
|
elif result > 4:
|
|
score += 75
|
|
elif result > 3:
|
|
score += 50
|
|
elif result > 2:
|
|
score += 25
|
|
|
|
return -score
|
|
|
|
varbound = np.array([[0,10000]]*3)
|
|
|
|
model = ga(function=f, dimension=3, variable_type='int', variable_boundaries=varbound)
|
|
|
|
model.run()
|