def straightSampleTrainTestVal(data): # number of games in data set N = len(data.X) # round up counts for train and validate Ntrain = int(N*0.6) + 1 Nval = int(N*0.8) + 1 # initialize train, test, and validate sets train = dataSet() test = dataSet() validate = dataSet() # get variable sets trainVars = vars(train).values() testVars = vars(test).values() valVars = vars(validate).values() dataVars = vars(data).values() # cycle through all variables in the dataset for i, var in enumerate(dataVars): # copy appropriate variables trainVars[i] += var[:Ntrain] valVars[i] += var[Ntrain:Nval] testVars[i] += var[Nval:] return train, test, validate
def randomSampleTrainTestVal(data): # number of games in data set, sample initial indexes N = len(data.X) samples = random.sample(range(N),N) # round up counts for train and validate Ntrain = int(N*0.6) + 1 Nval = int(N*0.8) + 1 # initialize train, test, and validate sets train = dataSet() test = dataSet() validate = dataSet() # get variable sets trainVars = vars(train).values() testVars = vars(test).values() valVars = vars(validate).values() dataVars = vars(data).values() # cycle through all variables in the dataset for i, var in enumerate(dataVars): mangled_var = var[samples] # copy appropriate variables trainVars[i] += mangled_var[:Ntrain] valVars[i] += mangled_var[Ntrain:Nval] testVars[i] += mangled_var[Nval:] return train, test, validate
def randomSampleTrainTestVal(data): # number of games in data set, sample initial indexes N = len(data.X) samples = random.sample(range(N), N) # round up counts for train and validate Ntrain = int(N * 0.6) + 1 Nval = int(N * 0.8) + 1 # initialize train, test, and validate sets train = dataSet() test = dataSet() validate = dataSet() # get variable sets trainVars = vars(train).values() testVars = vars(test).values() valVars = vars(validate).values() dataVars = vars(data).values() # cycle through all variables in the dataset for i, var in enumerate(dataVars): mangled_var = var[samples] # copy appropriate variables trainVars[i] += mangled_var[:Ntrain] valVars[i] += mangled_var[Ntrain:Nval] testVars[i] += mangled_var[Nval:] return train, test, validate
def straightSampleTrainTestVal(data): # number of games in data set N = len(data.X) # round up counts for train and validate Ntrain = int(N * 0.6) + 1 Nval = int(N * 0.8) + 1 # initialize train, test, and validate sets train = dataSet() test = dataSet() validate = dataSet() # get variable sets trainVars = vars(train).values() testVars = vars(test).values() valVars = vars(validate).values() dataVars = vars(data).values() # cycle through all variables in the dataset for i, var in enumerate(dataVars): # copy appropriate variables trainVars[i] += var[:Ntrain] valVars[i] += var[Ntrain:Nval] testVars[i] += var[Nval:] return train, test, validate
def sampleTrainTestValByYear(data): # number of games in data set N = len(data.X) # initialize train, test, and validate sets train = dataSet() test = dataSet() validate = dataSet() # get variable sets trainVars = vars(train).values() testVars = vars(test).values() valVars = vars(validate).values() dataVars = vars(data).values() # cycle through all games for n in range(N): # check year for training if data.yearList[n] < 2012: # copy all variables from that game to training set for i, var in enumerate(dataVars): trainVars[i].append(var[n]) # check year for test set (2013) elif data.yearList[n] == 2013: # copy all variables from that game to validation set for i, var in enumerate(dataVars): testVars[i].append(var[n]) # remaining data is validation set else: # copy all variables from that game to test set for i, var in enumerate(dataVars): valVars[i].append(var[n]) return train, test, validate