aialenti

['Tokyo', 'New York', 'Mexico City', 'Mumbai', 'São Paulo', 'Delhi',
       'Shanghai', 'Kolkata', 'Los Angeles', 'Dhaka', 'Buenos Aires',
       'Karachi', 'Cairo', 'Rio de Janeiro', 'Ōsaka', 'Beijing', 'Manila',
       'Moscow', 'Istanbul', 'Paris', 'Seoul', 'Lagos', 'Jakarta',
       'Guangzhou', 'Chicago', 'London', 'Lima', 'Tehran', 'Kinshasa',
       'Bogotá', 'Shenzhen', 'Wuhan', 'Hong Kong', 'Tianjin', 'Chennai',
       'Taipei', 'Bengalūru', 'Bangkok', 'Lahore', 'Chongqing', 'Miami',
       'Hyderabad', 'Dallas', 'Santiago', 'Philadelphia',
       'Belo Horizonte', 'Madrid', 'Houston', 'Ahmadābād',
       'Ho Chi Minh City', 'Washington', 'Atlanta', 'Toronto',

aialenti

['Tokyo', 'New York', 'Mexico City', 'Mumbai', 'São Paulo', 'Delhi',
       'Shanghai', 'Kolkata', 'Los Angeles', 'Dhaka', 'Buenos Aires',
       'Karachi', 'Cairo', 'Rio de Janeiro', 'Ōsaka', 'Beijing', 'Manila',
       'Moscow', 'Istanbul', 'Paris', 'Seoul', 'Lagos', 'Jakarta',
       'Guangzhou', 'Chicago', 'London', 'Lima', 'Tehran', 'Kinshasa',
       'Bogotá', 'Shenzhen', 'Wuhan', 'Hong Kong', 'Tianjin', 'Chennai',
       'Taipei', 'Bengalūru', 'Bangkok', 'Lahore', 'Chongqing', 'Miami',
       'Hyderabad', 'Dallas', 'Santiago', 'Philadelphia',
       'Belo Horizonte', 'Madrid', 'Houston', 'Ahmadābād',
       'Ho Chi Minh City', 'Washington', 'Atlanta', 'Toronto',

aialenti

class TSP():
    cities = None
    santa = None
    variables_dict = None
    x = None
    path = None
    sec_constraints = 0
    execution_time = 0

aialenti

aialenti

def preprocessing():
    # Read input data
    train = pd.read_csv("train.csv")
    categorical = ["X0", "X1", "X2", "X3", "X4", "X5", "X6", "X8"]

    #   Convert categorical data
    for c in categorical:
        group_by = train.groupby(by=c)["y"].mean().reset_index().rename(columns={"y": "{}_converted".format(c)})
        train = pd.merge(train, group_by, how='inner', on=c)

aialenti

param_grid = {'learning_rate': [float(v) for v in np.arange(0.01, 0.25, 0.01)],
              'colsample_bytree': [float(v) for v in np.arange(0.8, 1.01, 0.1)],
              'subsample': [float(v) for v in np.arange(0.5, 1.01, 0.1)],
              'n_estimators': [int(v) for v in np.arange(100, 3000, 100)],
              'reg_alpha': [float(v) for v in np.arange(0.01, 0.5, 0.05)],
              'max_depth': [int(v) for v in np.arange(3, 14, 1)],
              'gamma': [int(v) for v in np.arange(0, 10, 2)]
              }

aialenti

def get_grid_iterable():
    param_grid = {'learning_rate': [float(v) for v in np.arange(0.01, 0.25, 0.01)],
                  'colsample_bytree': [float(v) for v in np.arange(0.8, 1.01, 0.1)],
                  'subsample': [float(v) for v in np.arange(0.5, 1.01, 0.1)],
                  'n_estimators': [int(v) for v in np.arange(100, 3000, 100)],
                  'reg_alpha': [float(v) for v in np.arange(0.01, 0.5, 0.05)],
                  'max_depth': [int(v) for v in np.arange(3, 14, 1)],
                  'gamma': [int(v) for v in np.arange(0, 10, 2)]
                  }
    grid_iter = []

aialenti

    #   Read input data
    X, y = preprocessing()

    #   Create dataframe to collect the results
    tests_columns = ["test_nr", "cv_mean", "cv_min", "cv_max", "cv_median", "params"]
    test_id = 0
    tests = pd.DataFrame(columns=tests_columns)

aialenti

from sklearn.metrics import r2_score

def xgb_r2_score(preds, dtrain):
    labels = dtrain.get_label()
    return 'r2', r2_score(labels, preds)

aialenti

# Imports
import numpy as np
import pandas as pd
import xgboost as xgb
from interruptingcow import timeout
from sklearn.model_selection import KFold  # import KFold
from sklearn.metrics import r2_score
import json
from preprocessing import preprocessing, xgb_r2_score # The preprocessing and the r2 evaluation
from generate_grid import get_grid_iterable # The grid