Import libraries

--

import numpy as np
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
%matplotlib inline
import warnings
warnings.filterwarnings('ignore')

pd.set_option('display.max_rows', 20)
pd.set_option('display.max_columns', 500)
# pd.set_option('display.width', 1000)
data_folder = './data-for-code/'

# import the CSV as a pandas dataframe
df = pd.read_csv( data_folder + 'food_subgroup_intake_and_average_survival.csv')


# show the first five rows
df.shape
df.head(5)
png
df.shape, df.columns((108768, 51),
Index(['participant_id', 'participant_age', 'gender', 'food_group_name',
'food_subgroup_name', 'avg_food_weight_in_gms', 'acr', 'kidney_failed',
'systolic_pressure', 'diastolic_pressure', 'calorie', 'protein',
'sodium', 'carbohydrate', 'sugar', 'fibre', 'fat', 'saturated_fat',
'mono_fat', 'poly_fat', 'cholesterol', 'calcium', 'phosphorous',
'magnesium', 'potassium', 'alcohol', 'm_food_group_name',
'a_sample_food_code', 'a_sample_food', 'a_sample_food_name',
'food_group_id', 'albumin_urine_mu_g', 'albumin_urine_mg',
'creatinine_mu_mol', 'creatinine_mg', 'received_dialysis_in_12_months',
'kidney_stones', 'passed_kidney_stones_12_months',
'urinary_leakage_frequency', 'urine_lose_each_time',
'leak_during_activities', 'how_frequent_leak_occurs',
'urinated_before_reaching_toilet', 'how_frequent',
'leak_during_nonphysical_activities', 'how_frequest_leak_nonphysical',
'how_much_leak_bothering', 'how_much_daily_activities_affected',
'count_night_time_urinate', '2015', 'Average'],
dtype='object'))
df_food_subgroup_only = df[ ['participant_id', 'food_subgroup_name', 'avg_food_weight_in_gms'] ]
df_food_subgroup_only.head()

df_food_subgroup_only_index = df_food_subgroup_only.set_index(['participant_id', 'food_subgroup_name'])
df_food_subgroup_only_index

df_food_subgroup_only_index_unstack = df_food_subgroup_only_index.unstack()
df_food_subgroup_only_index_unstack.head()
png
df_without_food_subgroups = df.drop(['food_subgroup_name',  'avg_food_weight_in_gms'], axis=1)
df_without_food_subgroups.head()

df_without_food_subgroups_one_entry = df_without_food_subgroups.groupby(['participant_id']).mean()
df_without_food_subgroups_one_entry.head()
png
merged = df_food_subgroup_only_index_unstack.merge(df_without_food_subgroups_one_entry,  how='inner', left_on='participant_id', right_on='participant_id', left_index=True)
merged.head()
merged.shape
merged.head()
png
# remove extra spaces in the column names
column_dict = {}
for aName in merged.columns:
#print(aName)
try:
#if ( aName.index( 'avg_food_weight_in_gms' ) >= 0):
if 'avg_food_weight_in_gms' in aName:
#aName =
column_dict[aName] = aName[1] #aName.strip()
else:
column_dict[aName] = aName #.strip()

except:
print(aName)
continue

# column_dict
merged = merged.rename(index=str, columns=column_dict)
merged.head()
png
merged_f = merged.fillna(0)
merged_f.to_csv(data_folder + 'survival-food-subgroups-subgroup-names-on-columns-survival.csv')
merged_f.head()
df = merged_f
st = df
# Take only the important columns"""
st = df[

[
'Fruits and juices baby food', 'Nuts, Seeds, and Soy Products', 'Oils',
'Dried fruits',
'Added Sugars/Sugars and sweets',
'Alcoholic beverages', 'Citrus fruits, juices', 'Dark-green vegetables',
'Fruit juices and nectars excluding citrus',
'Meat, Poultry and Eggs subgroup', 'Milk desserts, sauces, gravies',
'Milks and milk drinks ', 'Nonalcoholic beverages', 'Other fruits',
'Other vegetables', 'Protein', 'Red and orange vegetables',
'Refined grains', 'Seafood',
'Solid Fats',
'Starchy vegetables', 'Water, noncarbonated', 'Whole grains',
'acr', 'acr_category', 'acr_category_2'

]
]
"""
"\nst = df[\n \n [\n 'Fruits and juices baby food', 'Nuts, Seeds, and Soy Products', 'Oils',\n 'Dried fruits', \n 'Added Sugars/Sugars and sweets',\n 'Alcoholic beverages', 'Citrus fruits, juices', 'Dark-green vegetables',\n 'Fruit juices and nectars excluding citrus',\n 'Meat, Poultry and Eggs subgroup', 'Milk desserts, sauces, gravies',\n 'Milks and milk drinks ', 'Nonalcoholic beverages', 'Other fruits',\n 'Other vegetables', 'Protein', 'Red and orange vegetables',\n 'Refined grains', 'Seafood', \n 'Solid Fats',\n 'Starchy vegetables', 'Water, noncarbonated', 'Whole grains',\n 'acr', 'acr_category', 'acr_category_2'\n \n ]\n]\n"# set palette : colorblind friendly# deep, muted, pastel, bright, dark, and colorblind.
current_palette = sns.color_palette()
sns.palplot(current_palette)
sns.set_palette('colorblind')
current_palette = sns.color_palette()
sns.palplot(current_palette)
png
png

Apply Machine Learning

Linear Regression

Bayesian

RandomForest Regression

st.columnsIndex(['Alcoholic beverages ', 'Beef', 'Cakes, cookies, pies, pastries, bars ',
'Cereals, not cooked or NS as to cooked ', 'Cheeses ',
'Citrus fruits, juices ',
'Crackers and salty snacks from grain products ',
'Creams and cream substitutes ', 'Dark-green vegetables ',
'Deep-yellow vegetables ', 'Dried fruits ', 'Egg mixtures',
'Egg substitutes ', 'Eggs', 'Fats', 'Fish and shellfish ',
'Formulated nutrition beverages, energy drinks, sports drinks, functional beverages ',
'Frankfurters, sausages, lunchmeats, meat spreads',
'Frozen or shelf-stable plate meals with meat, poultry, fish as major ingredient ',
'Fruit juices and nectars excluding citrus ',
'Fruits and juices baby food ',
'Grain mixtures, frozen plate meals, soups ',
'Gravies from meat, poultry, fish base ',
'Lamb, veal, game, other carcass meat ', 'Legumes ',
'Meat substitutes, mainly cereal protein ', 'Meat, NS as to type ',
'Meat, poultry, fish in gravy or sauce or creamed ',
'Meat, poultry, fish with starch item (includes white potatoes) ',
'Meat, poultry, fish with starch item and vegetables ',
'Meat, poultry, fish with vegetables (excluding white potatoes) ',
'Milk desserts, sauces, gravies ', 'Milks and milk drinks ',
'Mixtures mostly vegetables without meat, poultry, fish ',
'Nonalcoholic beverages ', 'Nuts, nut butters, and nut mixtures ',
'Oils ', 'Organ meats and mixtures ', 'Other fruits ',
'Other vegetables ',
'Pancakes, waffles, French toast, other grain products ',
'Pastas, cooked cereals, rice ', 'Pork', 'Poultry', 'Quick breads ',
'Salad dressings ', 'Sandwiches with meat, poultry, fish ',
'Seeds and seed mixtures ',
'Soups, broths, extracts from meat, poultry, fish base ',
'Sugars and sweets ', 'Tomatoes and tomato mixtures ',
'Vegetables and mixtures mostly vegetables baby food ',
'Vegetables with meat, poultry, fish ', 'Water, noncarbonated ',
'White potatoes and Puerto Rican starchy vegetables ',
'Yeast breads, rolls ', 'participant_age', 'gender', 'acr',
'kidney_failed', 'systolic_pressure', 'diastolic_pressure', 'calorie',
'protein', 'sodium', 'carbohydrate', 'sugar', 'fibre', 'fat',
'saturated_fat', 'mono_fat', 'poly_fat', 'cholesterol', 'calcium',
'phosphorous', 'magnesium', 'potassium', 'alcohol',
'a_sample_food_code', 'food_group_id', 'albumin_urine_mu_g',
'albumin_urine_mg', 'creatinine_mu_mol', 'creatinine_mg',
'received_dialysis_in_12_months', 'kidney_stones',
'passed_kidney_stones_12_months', 'urinary_leakage_frequency',
'urine_lose_each_time', 'leak_during_activities',
'how_frequent_leak_occurs', 'urinated_before_reaching_toilet',
'how_frequent', 'leak_during_nonphysical_activities',
'how_frequest_leak_nonphysical', 'how_much_leak_bothering',
'how_much_daily_activities_affected', 'count_night_time_urinate',
'2015', 'Average'],
dtype='object')
# as CKD data is between 10 to 15%, data need to be evenly distributed otherwise bias will be there
# alternatively, Five fold cross validation can be used
# Split data into Train and Test
# Use 10% of dataset as testing data
from sklearn.model_selection import train_test_split

Using 2015 survival probabilities as the target

Finding MSE, RMSE, Accuracy error

y = st['2015']
X = st.drop(columns=['Average', '2015'])

X_train, X_test, y_train, y_test = train_test_split( X, y, test_size=0.10, random_state=42)
len(y_train), len(y_test)

from sklearn.linear_model import LinearRegression
from sklearn.metrics import r2_score, mean_squared_error

# create the model
model = LinearRegression().fit(X_train, y_train)

train_predicted = model.predict(X_train)
test_predicted = model.predict(X_test)

train_mse = mean_squared_error(y_train, train_predicted)
test_mse = mean_squared_error(y_test, test_predicted)

print('MSE train data, MSE test data', train_mse, test_mse)
print('RMSE train data, RMSE test data', np.sqrt(np.absolute(train_mse)), np.sqrt(np.absolute(test_mse)) )
print('R2 train data, R2 test data', r2_score(y_train, train_predicted), r2_score(y_test, test_predicted))
list(abs(test_predicted) >0.5)
MSE train data, MSE test data 9.764225149069757 10.059648696067002
RMSE train data, RMSE test data 3.1247760158241356 3.171694924810235
R2 train data, R2 test data 0.8792316134983882 0.8696534568774607





[True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True,
True]

Use 2015 survival as the Target variable

# y = st['Average']
y = st['2015']
X = st.drop(columns=['Average', '2015'])

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.10, random_state=42)
len(y_train), len(y_test)

# create the model
model = LinearRegression().fit(X_train, y_train)

train_predicted = model.predict(X_train)
test_predicted = model.predict(X_test)

train_mse = mean_squared_error(y_train, train_predicted)
test_mse = mean_squared_error(y_test, test_predicted)

print('MSE train data, MSE test data', train_mse, test_mse)
print('RMSE train data, RMSE test data', np.sqrt(np.absolute(train_mse)), np.sqrt(np.absolute(test_mse)) )
print('R2 train data, R2 test data', r2_score(y_train, train_predicted), r2_score(y_test, test_predicted))
MSE train data, MSE test data 9.764225149066585 10.059648696069587
RMSE train data, RMSE test data 3.1247760158236275 3.1716949248106423
R2 train data, R2 test data 0.8792316134984274 0.8696534568774271

use cross validations using Linear Regression

2015 survival as the target

from sklearn.model_selection import cross_val_score


y = st['Average']
y = st['2015']
X = st.drop(columns=['Average', '2015'])
linear_regression_cross_validation_scores = cross_val_score(LinearRegression(), X, y, cv = 10)

print("Accuracy, Standard Deviations (+/- 2) :", linear_regression_cross_validation_scores.mean(), linear_regression_cross_validation_scores.std() * 2)
print('All Scores', linear_regression_cross_validation_scores)
print('\n\n')

# https://scikit-learn.org/stable/modules/generated/sklearn.metrics.accuracy_score.html
# " In multilabel classification, this function computes subset accuracy: the set of labels predicted for a sample must exactly match the corresponding set of labels in y_true. "
Data Not Normalized, Absolute ACR as the target
Accuracy, Standard Deviations (+/- 2) : 0.8737254002737421 0.009259078503739183
All Scores [0.87056245 0.87844429 0.87774739 0.87966261 0.87670929 0.86382014
0.87395519 0.86907356 0.87266426 0.87461483]

Apply Polynomial Regression

from sklearn.preprocessing import PolynomialFeatures


y = st['Average']
y = st['2015']
X = st.drop(columns=['Average', '2015'])
X_poly = PolynomialFeatures(degree = 2).fit_transform(X)

X_train, X_test, y_train, y_test = train_test_split(
X_poly, y, test_size=0.10, random_state=42)
Data Not Normalized, Absolute ACR as the targetpoly_regression = LinearRegression().fit(X_train, y_train)

poly_regression_train_pred = poly_regression.predict(X_train)
poly_regression_test_pred = poly_regression.predict(X_test)

poly_regression_train_mse = mean_squared_error(y_train, poly_regression_train_pred)
poly_regression_test_mse = mean_squared_error(y_test, poly_regression_test_pred)


print('MSE train data, MSE test data', poly_regression_train_mse, poly_regression_test_mse)
print('RMSE train data, RMSE test data', np.sqrt(np.absolute(poly_regression_train_mse)), np.sqrt(np.absolute(poly_regression_train_mse)))
print('R2 train data:, R2 test data', r2_score(y_train, poly_regression_train_pred), r2_score(y_test, poly_regression_test_pred))
MSE train data, MSE test data 1.0603702656911438 121.72287338237444
RMSE train data, RMSE test data 1.0297428153141657 1.0297428153141657
R2 train data:, R2 test data 0.9868848573105665 -0.5772077379340517
poly_regression_cv = cross_val_score(LinearRegression(), X_poly, y, cv = 5)
print("Accuracy and Standard Deviations (+/- 2) ", poly_regression_cv.mean(), poly_regression_cv.std() * 2)
print ('All Scores', poly_regression_cv)
Accuracy and Standard Deviations (+/- 2) -1.9086199777229425 2.9053185684926532
All Scores [-1.04165312 -0.50039128 -4.41478429 -2.66170411 -0.92456709]

RandomForestRegressor

from sklearn.ensemble import RandomForestRegressor


y = st['Average']
y = st['2015']
X = st.drop(columns=['Average', '2015'])
X_poly = PolynomialFeatures(degree = 2).fit_transform(X)

X_train, X_test, y_train, y_test = train_test_split(
X_poly, y, test_size=0.10, random_state=42)

# lees number of estimators can be tried to see the accuracy with less execution time
random_forest = RandomForestRegressor(n_estimators = 100, bootstrap=True, criterion='mse', max_depth=2).fit(X_train, y_train)

random_forest_train_pred = random_forest.predict(X_train)
random_forest_test_pred = random_forest.predict(X_test)

random_forest_train_mse = mean_squared_error(y_train, random_forest_train_pred)
random_forest_test_mse = mean_squared_error(y_test, random_forest_test_pred)

print('MSE train data, MSE test data', random_forest_train_mse, random_forest_test_mse)
print('RMSE train data, RMSE test data', np.sqrt(np.absolute(random_forest_train_mse)), np.sqrt(np.absolute(random_forest_train_mse)))
print('R2 train data, R2 test data', r2_score(y_train, random_forest_train_pred), r2_score(y_test, random_forest_test_pred))
Data Not Normalized, Absolute ACR as the target
MSE train data, MSE test data 6.505972950160663 6.445378515893091
RMSE train data, RMSE test data 2.550680879718328 2.550680879718328
R2 train data, R2 test data 0.9195311615803031 0.9164848759587985

RandomForestRegressor with cross validation

K Fold

random_forest_cv = cross_val_score(RandomForestRegressor(n_estimators = 100, bootstrap=True, criterion='mse', max_depth=2), X_poly, y, cv = 5)
print("Accuracy: Mean and Standard Deviations", random_forest_cv.mean(), random_forest_cv.std() * 2)
print('All scores', random_forest_cv)
Accuracy: Mean and Standard Deviations 0.9199252729983216 0.013099232983670599
All scores [0.91691393 0.93134251 0.91404089 0.91434822 0.92298083]

From: https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestRegressor.html Ref: Example Parameters:
RandomForestRegressor(bootstrap=True, criterion=’mse’, max_depth=2, max_features=’auto’, max_leaf_nodes=None, min_impurity_decrease=0.0, min_impurity_split=None, min_samples_leaf=1, min_samples_split=2, min_weight_fraction_leaf=0.0, n_estimators=100, n_jobs=None, oob_score=False, random_state=0, verbose=0, warm_start=False)

random_forest_cvarray([0.91691393, 0.93134251, 0.91404089, 0.91434822, 0.92298083])

Bayesian

from sklearn.linear_model import BayesianRidge


y = st['Average']
y = st['2015']
X = st.drop(columns=['Average', '2015'])

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.10, random_state=42)

bayesian = BayesianRidge().fit(X_train, y_train)

bayesian_train_pred = bayesian.predict(X_train)
bayesian_test_pred = bayesian.predict(X_test)

bayesian_train_mse = mean_squared_error(y_train, bayesian_train_pred)
bayesian_test_mse = mean_squared_error(y_test, bayesian_test_pred)

print('MSE train data, MSE test data', bayesian_train_mse, bayesian_test_mse)
print('RMSE train data, RMSE test data', np.sqrt(np.absolute(bayesian_train_mse)), np.sqrt(np.absolute(bayesian_train_mse)))

print('R2 train data, R2 test data', r2_score(y_train, bayesian_train_pred), r2_score(y_test, bayesian_test_pred))

print('#########################################')
MSE train data, MSE test data 9.794079101552265 10.061498052597019
RMSE train data, RMSE test data 3.1295493448022618 3.1295493448022618
R2 train data, R2 test data 0.8788623662087196 0.8696294940892986
#########################################

Polynomial

#X_poly =  PolynomialFeatures(degree = 2).fit_transform(X)
X_train, X_test, y_train, y_test = train_test_split(X_poly, y, test_size=0.10, random_state=42)

bayesian = BayesianRidge().fit(X_train, y_train)

bayesian_train_pred = bayesian.predict(X_train)
bayesian_test_pred = bayesian.predict(X_test)

bayesian_train_mse = mean_squared_error(y_train, bayesian_train_pred)
bayesian_test_mse = mean_squared_error(y_test, bayesian_test_pred)

print('Bayesian on Polynomial fit')
print('MSE train data, MSE test data', bayesian_train_mse, bayesian_test_mse)
print('RMSE train data, RMSE test data', np.sqrt(np.absolute(bayesian_train_mse)), np.sqrt(np.absolute(bayesian_train_mse)))

print('R2 train data, R2 test data', r2_score(y_train, bayesian_train_pred), r2_score(y_test, bayesian_test_pred))
Bayesian on Polynomial fit
MSE train data, MSE test data 4.208880026402317 10.302807177360028
RMSE train data, RMSE test data 2.0515555138485326 2.0515555138485326
R2 train data, R2 test data 0.94794265371729 0.8665027636052528

Bayesian with Cross Validation

Polynomial X is used

bayesian_cv = cross_val_score(BayesianRidge(), X_poly, y, cv = 5)
print("Accuracy and Standard Deviations", bayesian_cv.mean(), bayesian_cv.std() * 2)
print('All Scores', bayesian_cv)
Accuracy and Standard Deviations 0.764463816301802 0.3019204129703462
All Scores [0.46550975 0.86175398 0.85979255 0.83113807 0.80412473]

https://machinelearningmastery.com/metrics-evaluate-machine-learning-algorithms-python/

“Some evaluation metrics (like mean squared error) are naturally descending scores (the smallest score is best) and as such are reported as negative by the cross_val_score() function. This is important to note, because some scores will be reported as negative that by definition can never be negative.”

The following can be ignored, Kfold cross validation is already considered above

References

Projects mentioned on: http://sitestree.com/prediction-bayesian-regression-concepts-example-projects/

Insurance HealthCare Costs: https://github.com/techshot25/HealthCare Linear and Bayesian modeling in R: Predicting movie popularity https://towardsdatascience.com/linear-and-bayesian-modelling-in-r-predicting-movie-popularity-6c8ef0a44184

Bayesian-Stock-Price-Prediction https://github.com/lschlessinger1/Bayesian-Stock-Price-Prediction

Bayesian Prediction: Well (Oil) Production https://github.com/jpgrana/bayesian-approach-predicting-well-production

Binary Classification on Stock Market (S&P 500) using Naive Bayes and Logistic Regression https://github.com/NeilPrabhu/Stock-Prediction

Naive Bayes Weather Prediction https://github.com/husnainfareed/simple-naive-bayes-weather-prediction/blob/master/bayes.py

Regression Predict Fuel Efficiency: https://www.tensorflow.org/tutorials/keras/basic_regression

Regression-Example-Predicting-House-Prices https://github.com/andersy005/deep-learning/blob/master/keras/04-A-Regression-Example-Predicting-House-Prices.ipynb

Stock Price Prediction using Regression https://github.com/chaitjo/regression-stock-prediction

Concept: Predicting the Future with Bayes’ Theorem https://fs.blog/2018/09/bayes-theorem/

Chapter 5 — Bayesian Prediction https://www.sciencedirect.com/science/article/pii/B9780123748546000089

Books:

Bayesian Methods for Hackers https://github.com/CamDavidsonPilon/Probabilistic-Programming-and-Bayesian-Methods-for-Hackers

Multiple-linear-regression https://github.com/topics/multiple-linear-regression

Making Predictions with Regression Analysis https://statisticsbyjim.com/regression/predictions-regression/

Regression and Prediction http://jukebox.esc13.net/untdeveloper/RM/Stats_Module_5/Stats_Module_56.html

--

--

Justetc Social Services (non-profit)
Data Science Project Development

All proceeds from Medium will go to Justetc Social Services ( non-profit). Justetc Social Services provides services in the Training and Education Areas.