From 2b953583e27e9f8722abbd4da794e66605cc9b54 Mon Sep 17 00:00:00 2001 From: Tanushree Tunstall Date: Sat, 3 Sep 2022 12:28:36 +0100 Subject: [PATCH] added combined model FS code and run script --- scripts/ml/boruta_test_clfs.py | 121 +++++++++ scripts/ml/combined_model/cm_logo_skf_FS.py | 280 ++++++++++++++++++++ scripts/ml/combined_model/run_cm_logo.py | 32 +++ scripts/ml/combined_model/run_cm_logo_FS.py | 204 ++++++++++++++ scripts/ml/untitled5.py | 52 ++++ scripts/ml/untitled6.py | 136 ++++++++++ scripts/ml/untitled7_boruta.py | 221 +++++++++++++++ 7 files changed, 1046 insertions(+) create mode 100644 scripts/ml/boruta_test_clfs.py create mode 100755 scripts/ml/combined_model/cm_logo_skf_FS.py create mode 100644 scripts/ml/combined_model/run_cm_logo.py create mode 100644 scripts/ml/combined_model/run_cm_logo_FS.py create mode 100644 scripts/ml/untitled5.py create mode 100644 scripts/ml/untitled6.py create mode 100644 scripts/ml/untitled7_boruta.py diff --git a/scripts/ml/boruta_test_clfs.py b/scripts/ml/boruta_test_clfs.py new file mode 100644 index 0000000..fee80e6 --- /dev/null +++ b/scripts/ml/boruta_test_clfs.py @@ -0,0 +1,121 @@ +#!/usr/bin/env python3 +# -*- coding: utf-8 -*- +""" +Created on Fri Sep 2 16:10:44 2022 + +@author: tanu +""" +from sklearn.ensemble import VotingClassifier +from sklearn.ensemble import BaggingClassifier +from sklearn.ensemble import AdaBoostClassifier, RandomForestClassifier, ExtraTreesClassifier +from boruta import BorutaPy + +fooD = combined_DF_OS(combined_df) + +numerical_ix = fooD['X'].select_dtypes(include=['int64', 'float64']).columns +numerical_ix +print("\nNo. of numerical indices:", len(numerical_ix)) + +categorical_ix = fooD['X'].select_dtypes(include=['object', 'bool']).columns +categorical_ix +print("\nNo. of categorical indices:", len(categorical_ix)) + + +var_type = "mixed" + +if var_type == 'mixed': + + t = [('num', MinMaxScaler(), numerical_ix) + , ('cat', OneHotEncoder(), categorical_ix)] + +col_transform = ColumnTransformer(transformers = t + , remainder='passthrough') +#--------------ALEX help +# col_transform +# col_transform.fit(X) +# test = col_transform.transform(X) +# print(col_transform.get_feature_names_out()) + +# foo = col_transform.fit_transform(X) +Xm_train = col_transform.fit_transform(fooD['X']) +fooD['X'].shape +Xm_train.shape + +Xm_test = col_transform.fit_transform(fooD['X_bts']) +fooD['X_bts'].shape +Xm_test.shape + +X_train = Xm_train.copy() +X_test = Xm_test.copy() +X_train.shape +X_test.shape + +y_train = fooD['y'] +y_test = fooD['y_bts'] +y_train.shape +y_test.shape + +# perhaps +#col_transform.fit(fooD['X']) +#encoded_colnames = pd.Index(col_transform.get_feature_names_out()) +#====================== +# 1 model +n_jobs = os.cpu_count() +njobs = {'n_jobs': n_jobs } +rs = {'random_state': 42} + +rf_all_features = RandomForestClassifier(n_estimators=1000, max_depth=5 + , **rs, **njobs) + +#rf_all_features = VotingClassifier(estimators=1000) +rf_all_features = BaggingClassifier(random_state=1, n_estimators=100, verbose = 3, **njobs) +rf_all_features = AdaBoostClassifier(random_state=1, n_estimators=1000) +rf_all_features = ExtraTreesClassifier(random_state=1, n_estimators=1000, max_depth=5, verbose = 3) +rf_all_features = DecisionTreeClassifier(random_state=1, max_depth=5) + + +rf_all_features.fit(X_train, np.array(y_train)) +accuracy_score(y_test, rf_all_features.predict(X_test)) +matthews_corrcoef(y_test, rf_all_features.predict(X_test)) + +# BORUTA +boruta_selector = BorutaPy(rf_all_features,**rs, verbose = 3) +boruta_selector.fit(np.array(X_train), np.array(y_train)) + +# Tells you how many features: GOOD +print("Ranking: ", boruta_selector.ranking_) +print("No. of significant features: ", boruta_selector.n_features_) + + +cm_df = combined_df.drop(['gene_name', 'dst', 'dst_mode'], axis = 1) +col_transform.fit(cm_df) +col_transform.get_feature_names_out() + +var_type_colnames = col_transform.get_feature_names_out() +var_type_colnames = pd.Index(var_type_colnames) + +if var_type == 'mixed': + print('\nVariable type is:', var_type + , '\nNo. of columns in input_df:', len(cm_df.columns) + , '\nNo. of columns post one hot encoder:', len(var_type_colnames)) +else: + print('\nNo. of columns in input_df:', len(input_df.columns)) + + +selected_rf_features = pd.DataFrame({'Feature':list(var_type_colnames), + 'Ranking':boruta_selector.ranking_}) +sel_rf_features_sorted = selected_rf_features.sort_values(by='Ranking') + + +sel_features = var_type_colnames[boruta_selector.support_] + + +# tells you the ranking: GOOD +#foo2 = selected_rf_features.sort_values(by='Ranking') + +X_important_train = boruta_selector.transform(np.array(X_train)) +X_important_test = boruta_selector.transform(np.array(X_test)) + +rf_all_features.fit(X_important_train, y_train) +accuracy_score(y_test, rf_all_features.predict(X_important_test)) +matthews_corrcoef(y_test, rf_all_features.predict(X_important_test)) diff --git a/scripts/ml/combined_model/cm_logo_skf_FS.py b/scripts/ml/combined_model/cm_logo_skf_FS.py new file mode 100755 index 0000000..7a77efa --- /dev/null +++ b/scripts/ml/combined_model/cm_logo_skf_FS.py @@ -0,0 +1,280 @@ +#!/usr/bin/env python3 +# -*- coding: utf-8 -*- +""" +Created on Wed Jun 29 19:44:06 2022 + +@author: tanu +""" +import sys, os +import pandas as pd +import numpy as np +import re +from copy import deepcopy +from sklearn import linear_model +from sklearn import datasets +from collections import Counter + +from sklearn.linear_model import LogisticRegression, LogisticRegressionCV +from sklearn.linear_model import RidgeClassifier, RidgeClassifierCV, SGDClassifier, PassiveAggressiveClassifier + +from sklearn.naive_bayes import BernoulliNB +from sklearn.neighbors import KNeighborsClassifier +from sklearn.svm import SVC +from sklearn.tree import DecisionTreeClassifier, ExtraTreeClassifier +from sklearn.ensemble import RandomForestClassifier, ExtraTreesClassifier, AdaBoostClassifier, GradientBoostingClassifier, BaggingClassifier +from sklearn.naive_bayes import GaussianNB +from sklearn.gaussian_process import GaussianProcessClassifier, kernels +from sklearn.gaussian_process.kernels import RBF, DotProduct, Matern, RationalQuadratic, WhiteKernel + +from sklearn.discriminant_analysis import LinearDiscriminantAnalysis, QuadraticDiscriminantAnalysis +from sklearn.neural_network import MLPClassifier + +from sklearn.svm import SVC +from xgboost import XGBClassifier +from sklearn.naive_bayes import MultinomialNB +from sklearn.preprocessing import StandardScaler, MinMaxScaler, OneHotEncoder + +from sklearn.compose import ColumnTransformer +from sklearn.compose import make_column_transformer + +from sklearn.metrics import make_scorer, confusion_matrix, accuracy_score, balanced_accuracy_score, precision_score, average_precision_score, recall_score +from sklearn.metrics import roc_auc_score, roc_curve, f1_score, matthews_corrcoef, jaccard_score, classification_report + +# added +from sklearn.model_selection import train_test_split, cross_validate, cross_val_score, LeaveOneOut, KFold, RepeatedKFold, cross_val_predict + +from sklearn.model_selection import train_test_split, cross_validate, cross_val_score +from sklearn.model_selection import StratifiedKFold,RepeatedStratifiedKFold, RepeatedKFold + +from sklearn.pipeline import Pipeline, make_pipeline + +from sklearn.feature_selection import RFE, RFECV + +import itertools +import seaborn as sns +import matplotlib.pyplot as plt + +from statistics import mean, stdev, median, mode + +from imblearn.over_sampling import RandomOverSampler +from imblearn.under_sampling import RandomUnderSampler +from imblearn.over_sampling import SMOTE +from sklearn.datasets import make_classification +from imblearn.combine import SMOTEENN +from imblearn.combine import SMOTETomek + +from imblearn.over_sampling import SMOTENC +from imblearn.under_sampling import EditedNearestNeighbours +from imblearn.under_sampling import RepeatedEditedNearestNeighbours + +from sklearn.model_selection import GridSearchCV +from sklearn.base import BaseEstimator +from sklearn.impute import KNNImputer as KNN +import json +import argparse +import re +import itertools +from sklearn.model_selection import LeaveOneGroupOut +from boruta import BorutaPy + +############################################################################### +# homedir = os.path.expanduser("~") +# sys.path.append(homedir + '/git/LSHTM_analysis/scripts/ml/ml_functions') +# sys.path +############################################################################### +#outdir = "/home/tanu/git/LSHTM_ML/output/feature_selection/" + +#==================== +# Import ML functions +#==================== +#from ml_data_combined import * +#from MultClfs import * +#from GetMLData import * +#from SplitTTS import * + +#skf_cv = StratifiedKFold(n_splits = 10 , shuffle = True, random_state = 42) + +#logo = LeaveOneGroupOut() + +######################################################################## +# COMPLETE data: No tts_split +######################################################################## +#%% +def CMLogoSkf_FS(cm_input_df + , all_genes = ["embb", "katg", "rpob", "pnca", "gid", "alr"] + , bts_genes = ["embb", "katg", "rpob", "pnca", "gid"] + , cols_to_drop = ['dst', 'dst_mode', 'gene_name'] + , target_var = 'dst_mode' + , gene_group = 'gene_name' + , std_gene_omit = [] + , var_type = ['numerical', 'categorical','mixed'] + + ): + + n_jobs = os.cpu_count() + njobs = {'n_jobs': n_jobs } + rs = {'random_state': 42} + + cm_gene_featuresD = {} + for bts_gene in bts_genes: + print('\n BTS gene:', bts_gene) + if not std_gene_omit: + training_genesL = ['alr'] + else: + training_genesL = [] + + tr_gene_omit = std_gene_omit + [bts_gene] + n_tr_genes = (len(bts_genes) - (len(std_gene_omit))) + #n_total_genes = (len(bts_genes) - len(std_gene_omit)) + n_total_genes = len(all_genes) + + training_genesL = training_genesL + list(set(bts_genes) - set(tr_gene_omit)) + #training_genesL = [element for element in bts_genes if element not in tr_gene_omit] + + print('\nTotal genes: ', n_total_genes + ,'\nTraining on:', n_tr_genes + ,'\nTraining on genes:', training_genesL + , '\nOmitted genes:', tr_gene_omit + , '\nBlind test gene:', bts_gene) + + print('\nDim of data:', cm_input_df.shape) + + tts_split_type = "logo_skf_BT_" + bts_gene + + #------- + # training + #------ + cm_training_df = cm_input_df[~cm_input_df['gene_name'].isin(tr_gene_omit)] + + cm_X = cm_training_df.drop(cols_to_drop, axis=1, inplace=False) + #cm_y = cm_training_df.loc[:,'dst_mode'] + cm_y = cm_training_df.loc[:, target_var] + + gene_group = cm_training_df.loc[:,'gene_name'] + + print('\nTraining data dim:', cm_X.shape + , '\nTraining Target dim:', cm_y.shape) + + if all(cm_X.columns.isin(cols_to_drop) == False): + print('\nChecked training df does NOT have Target var') + else: + sys.exit('\nFAIL: training data contains Target var') + + #--------------- + # BTS: genes + #--------------- + cm_test_df = cm_input_df[cm_input_df['gene_name'].isin([bts_gene])] + + cm_bts_X = cm_test_df.drop(cols_to_drop, axis = 1, inplace = False) + #cm_bts_y = cm_test_df.loc[:, 'dst_mode'] + cm_bts_y = cm_test_df.loc[:, target_var] + + print('\nTEST data dim:' , cm_bts_X.shape + , '\nTEST Target dim:' , cm_bts_y.shape) + + print("Running Multiple models on LOGO with SKF") + + + # REASSIGN for simplicity + # X + X_train = cm_X.copy() + X_test = cm_bts_X.copy() + X_train.shape + X_test.shape + + # Y + y_train = cm_y.copy() + y_test = cm_bts_y.copy() + y_train.shape + y_test.shape + + +############################################################################## + #PREPROCESS + + numerical_ix = X_train.select_dtypes(include=['int64', 'float64']).columns + numerical_ix + print("\nNo. of numerical indices:", len(numerical_ix)) + + categorical_ix = X_train.select_dtypes(include=['object', 'bool']).columns + categorical_ix + print("\nNo. of categorical indices:", len(categorical_ix)) + + #====================================================== + # Determine preprocessing steps ~ var_type + #====================================================== + + if var_type == 'numerical': + t = [('num', MinMaxScaler(), numerical_ix)] + + if var_type == 'categorical': + t = [('cat', OneHotEncoder(), categorical_ix)] + + if var_type == 'mixed': + t = [('num', MinMaxScaler(), numerical_ix) + , ('cat', OneHotEncoder(), categorical_ix)] + + col_transform = ColumnTransformer(transformers = t + , remainder='passthrough') + + col_transform.fit(X_train) + col_transform.get_feature_names_out() + + var_type_colnames = col_transform.get_feature_names_out() + var_type_colnames = pd.Index(var_type_colnames) + + if var_type == 'mixed': + print('\nVariable type is:', var_type + , '\nNo. of columns in input_df:', len(X_train.columns) + , '\nNo. of columns post one hot encoder:', len(var_type_colnames)) + else: + print('\nNo. of columns in input_df:', len(cm_input_df.columns)) + + +############################################################################## + # FS: Random Forest + Boruta + + X_train = col_transform.fit_transform(X_train) + X_test = col_transform.fit_transform(X_test) + + fs_clf = "RandomForestClassifier" + rf_all_features = RandomForestClassifier(n_estimators=1000, max_depth=5 + , **rs, **njobs) + + # fit + rf_all_features.fit(np.array(X_train), np.array(y_train)) + print("RF, baseline MCC:", matthews_corrcoef(y_test, rf_all_features.predict(X_test))) + + # BORUTA and fit + boruta_selector = BorutaPy(rf_all_features,**rs, verbose = 3) + boruta_selector.fit(np.array(X_train), np.array(y_train)) + + # Get chosen features + print("Ranking: ", boruta_selector.ranking_) + print("No. of significant features: ", boruta_selector.n_features_) + + + X_important_train = boruta_selector.transform(np.array(X_train)) + X_important_test = boruta_selector.transform(np.array(X_test)) + + # just retesting with selected features on RF itselfs + rf_all_features.fit(X_important_train, y_train) + print("RF, Boruta MCC:", matthews_corrcoef(y_test, rf_all_features.predict(X_important_test))) + + selected_rf_features = pd.DataFrame({'Feature':list(var_type_colnames), + 'Ranking':boruta_selector.ranking_}) + + sel_rf_features_sorted = selected_rf_features.sort_values(by='Ranking') + + + sel_features = var_type_colnames[boruta_selector.support_] + cm_gene_featuresD.update({bts_gene: { + "sel_features": sel_features + , "fs_ranking" : sel_rf_features_sorted + , "fs_model_name": fs_clf + } + } + ) + + + return(cm_gene_featuresD) diff --git a/scripts/ml/combined_model/run_cm_logo.py b/scripts/ml/combined_model/run_cm_logo.py new file mode 100644 index 0000000..e3a0c1e --- /dev/null +++ b/scripts/ml/combined_model/run_cm_logo.py @@ -0,0 +1,32 @@ +#!/usr/bin/env python3 +# -*- coding: utf-8 -*- +""" +Created on Sat Sep 3 09:43:22 2022 + +@author: tanu +""" + +############################################################################### +homedir = os.path.expanduser("~") +sys.path.append(homedir + '/home/tanu/git/LSHTM_analysis/scripts/ml/combined_model') +sys.path.append(homedir + '/home/tanu/git/LSHTM_analysis/scripts/ml/ml_functions') +sys.path.append(homedir + '/home/tanu/git/LSHTM_analysis/scripts/ml') + +from MultClfs import * +############################################################################### + +#%% RUN: Combined model Baseline +outdir_cg = "/home/tanu/git/LSHTM_ML/output/combined/" +#=============== +# Complete Data +#=============== +CombinedModelML(cm_input_df = combined_df, outdir = outdir_cg, file_suffix = "complete") +CombinedModelML(cm_input_df = combined_df, outdir = outdir_cg, std_gene_omit=['alr'], file_suffix = "complete") + +#=============== +# Actual Data +#=============== +CombinedModelML(cm_input_df = combined_df_actual, outdir = outdir_cg, file_suffix = "actual") +CombinedModelML(cm_input_df = combined_df_actual, outdir = outdir_cg, std_gene_omit=['alr'], file_suffix = "actual") + + diff --git a/scripts/ml/combined_model/run_cm_logo_FS.py b/scripts/ml/combined_model/run_cm_logo_FS.py new file mode 100644 index 0000000..2479441 --- /dev/null +++ b/scripts/ml/combined_model/run_cm_logo_FS.py @@ -0,0 +1,204 @@ +#!/usr/bin/env python3 +# -*- coding: utf-8 -*- +""" +Created on Fri Sep 2 19:17:46 2022 + +@author: tanu +""" +############################################################################### +homedir = os.path.expanduser("~") +sys.path.append(homedir + '/home/tanu/git/LSHTM_analysis/scripts/ml/combined_model') +sys.path.append(homedir + '/home/tanu/git/LSHTM_analysis/scripts/ml/ml_functions') +sys.path.append(homedir + '/home/tanu/git/LSHTM_analysis/scripts/ml') + +from MultClfs import * +from cm_logo_skf_FS import * + +############################################################################### +#%% FS with all genes in training +############################################################################### + +# 1. Select Features +boruta_features = CMLogoSkf_FS(cm_input_df = combined_df,var_type = 'mixed', file_suffix = "complete") + +# 2. Find original column names of features +# if it starts with num__, get rid of num__ +# if it starts with cat__, get rid of cat__ and the _ at the end +for i in boruta_features: + print(i) + boruta_features[i]['sel_features']=[re.sub('^num__|cat__(.*)_\d*$',r'\1', current_thing) for current_thing in boruta_features[i]['sel_features']] + boruta_features[i]['sel_features'] = list(set(boruta_features[i]['sel_features'])) + +# write json +OutFile_6Tgenes = "/home/tanu/git/LSHTM_ML/output/feature_selection/boruta_features_6_Tgenes.json" +pd.DataFrame(boruta_features).to_json(path_or_buf=OutFile_6Tgenes) + +# 3. Run all classification models using original column names from (2) +combined_df_embb=combined_df[boruta_features['embb']['sel_features']+['dst', 'dst_mode', 'gene_name']] +combined_df_katg=combined_df[boruta_features['katg']['sel_features']+['dst', 'dst_mode', 'gene_name']] +combined_df_pnca=combined_df[boruta_features['pnca']['sel_features']+['dst', 'dst_mode', 'gene_name']] +combined_df_gid= combined_df[boruta_features['gid' ]['sel_features']+['dst', 'dst_mode', 'gene_name']] +combined_df_rpob= combined_df[boruta_features['rpob' ]['sel_features']+['dst', 'dst_mode', 'gene_name']] + + +# from /home/tanu/git/LSHTM_analysis/scripts/ml/ml_functions/MultClf.py +CombinedModelML(combined_df_embb + , all_genes = ["embb", "katg", "rpob", "pnca", "gid", "alr"] + , bts_genes = ["embb"] + , cols_to_drop = ['dst', 'dst_mode', 'gene_name'] + , target_var = 'dst_mode' + , gene_group = 'gene_name' + , std_gene_omit = [] + , output_dir = "/home/tanu/git/LSHTM_ML/output/feature_selection/" + , file_suffix = "FS" + ) + + +CombinedModelML(combined_df_katg + , all_genes = ["embb", "katg", "rpob", "pnca", "gid", "alr"] + , bts_genes = ["katg"] + , cols_to_drop = ['dst', 'dst_mode', 'gene_name'] + , target_var = 'dst_mode' + , gene_group = 'gene_name' + , std_gene_omit = [] + , output_dir = "/home/tanu/git/LSHTM_ML/output/feature_selection/" + , file_suffix = "FS" + ) + + +CombinedModelML(combined_df_pnca + , all_genes = ["embb", "katg", "rpob", "pnca", "gid", "alr"] + , bts_genes = ["pnca"] + , cols_to_drop = ['dst', 'dst_mode', 'gene_name'] + , target_var = 'dst_mode' + , gene_group = 'gene_name' + , std_gene_omit = [] + , output_dir = "/home/tanu/git/LSHTM_ML/output/feature_selection/" + , file_suffix = "FS" + ) + +CombinedModelML(combined_df_gid + , all_genes = ["embb", "katg", "rpob", "pnca", "gid", "alr"] + , bts_genes = ["gid"] + , cols_to_drop = ['dst', 'dst_mode', 'gene_name'] + , target_var = 'dst_mode' + , gene_group = 'gene_name' + , std_gene_omit = [] + , output_dir = "/home/tanu/git/LSHTM_ML/output/feature_selection/" + , file_suffix = "FS" + ) + +CombinedModelML(combined_df_rpob + , all_genes = ["embb", "katg", "rpob", "pnca", "gid", "alr"] + , bts_genes = ["rpob"] + , cols_to_drop = ['dst', 'dst_mode', 'gene_name'] + , target_var = 'dst_mode' + , gene_group = 'gene_name' + , std_gene_omit = [] + , output_dir = "/home/tanu/git/LSHTM_ML/output/feature_selection/" + , file_suffix = "FS" + ) + + +# write all feature rankings +for i in boruta_features: + print (i) + gene_fs_ranking = boruta_features[i]['fs_ranking'] + gene_fs_ranking.to_csv("/home/tanu/git/LSHTM_ML/output/feature_selection/"+ i + "_boruta_featues_6Tgenes.csv") + + +############################################################################### +#%% FS withour training including ALR +############################################################################### +# With training omitting alr +boruta_features_omit_alr = CMLogoSkf_FS(cm_input_df = combined_df + , std_gene_omit = ['alr'] + , var_type = 'mixed') + +# 2. Find original column names of features +# if it starts with num__, get rid of num__ +# if it starts with cat__, get rid of cat__ and the _ at the end +for i in boruta_features_omit_alr: + print(i) + boruta_features_omit_alr[i]['sel_features']=[re.sub('^num__|cat__(.*)_\d*$',r'\1', current_thing) for current_thing in boruta_features[i]['sel_features']] + boruta_features_omit_alr[i]['sel_features'] = list(set(boruta_features_omit_alr[i]['sel_features'])) + +# write json +OutFile_5Tgenes = "/home/tanu/git/LSHTM_ML/output/feature_selection/boruta_features_5_Tgenes.json" +pd.DataFrame(boruta_features_omit_alr).to_json(path_or_buf=OutFile_5Tgenes) + +# 3. Run all classification models using original column names from (2) +cm_input_df5 = combined_df[~combined_df['gene_name'].isin(omit_gene_alr)] + +combined_df_embb_no_alr = cm_input_df5[boruta_features_omit_alr['embb']['sel_features']+['dst', 'dst_mode', 'gene_name']] +combined_df_katg_no_alr = cm_input_df5[boruta_features_omit_alr['katg']['sel_features']+['dst', 'dst_mode', 'gene_name']] +combined_df_pnca_no_alr = cm_input_df5[boruta_features_omit_alr['pnca']['sel_features']+['dst', 'dst_mode', 'gene_name']] +combined_df_gid_no_alr = cm_input_df5[boruta_features_omit_alr['gid' ]['sel_features']+['dst', 'dst_mode', 'gene_name']] +combined_df_rpob_no_alr = cm_input_df5[boruta_features_omit_alr['rpob' ]['sel_features']+['dst', 'dst_mode', 'gene_name']] + + +CombinedModelML(combined_df_embb_no_alr + , all_genes = ["embb", "katg", "rpob", "pnca", "gid", "alr"] + , bts_genes = ["embb"] + , cols_to_drop = ['dst', 'dst_mode', 'gene_name'] + , target_var = 'dst_mode' + , gene_group = 'gene_name' + , std_gene_omit = ["alr"] + , output_dir = "/home/tanu/git/LSHTM_ML/output/feature_selection/" + , file_suffix = "FS_no_Talr" + ) + + +CombinedModelML(combined_df_katg_no_alr + , all_genes = ["embb", "katg", "rpob", "pnca", "gid", "alr"] + , bts_genes = ["katg"] + , cols_to_drop = ['dst', 'dst_mode', 'gene_name'] + , target_var = 'dst_mode' + , gene_group = 'gene_name' + , std_gene_omit = ["alr"] + , output_dir = "/home/tanu/git/LSHTM_ML/output/feature_selection/" + , file_suffix = "FS_no_Talr" + ) + + +CombinedModelML(combined_df_pnca_no_alr + , all_genes = ["embb", "katg", "rpob", "pnca", "gid", "alr"] + , bts_genes = ["pnca"] + , cols_to_drop = ['dst', 'dst_mode', 'gene_name'] + , target_var = 'dst_mode' + , gene_group = 'gene_name' + , std_gene_omit = ["alr"] + , output_dir = "/home/tanu/git/LSHTM_ML/output/feature_selection/" + , file_suffix = "FS_no_Talr" + ) + +CombinedModelML(combined_df_gid_no_alr + , all_genes = ["embb", "katg", "rpob", "pnca", "gid", "alr"] + , bts_genes = ["gid"] + , cols_to_drop = ['dst', 'dst_mode', 'gene_name'] + , target_var = 'dst_mode' + , gene_group = 'gene_name' + , std_gene_omit = ["alr"] + , output_dir = "/home/tanu/git/LSHTM_ML/output/feature_selection/" + , file_suffix = "FS_no_Talr" + ) + +CombinedModelML(combined_df_rpob_no_alr + , all_genes = ["embb", "katg", "rpob", "pnca", "gid", "alr"] + , bts_genes = ["rpob"] + , cols_to_drop = ['dst', 'dst_mode', 'gene_name'] + , target_var = 'dst_mode' + , gene_group = 'gene_name' + , std_gene_omit = ["alr"] + , output_dir = "/home/tanu/git/LSHTM_ML/output/feature_selection/" + , file_suffix = "FS_no_Talr" + ) + + +# write all feature rankings +for i in boruta_features_omit_alr: + print (i) + gene_fs_ranking_no_alr = boruta_features_omit_alr[i]['fs_ranking'] + gene_fs_ranking_no_alr.to_csv("/home/tanu/git/LSHTM_ML/output/feature_selection/"+ i + "_boruta_featues_5Tgenes.csv") + + diff --git a/scripts/ml/untitled5.py b/scripts/ml/untitled5.py new file mode 100644 index 0000000..b2090b9 --- /dev/null +++ b/scripts/ml/untitled5.py @@ -0,0 +1,52 @@ +#!/usr/bin/env python3 +# -*- coding: utf-8 -*- +""" +Created on Fri Sep 2 11:11:49 2022 + +@author: tanu +""" +# https://towardsdatascience.com/explain-feature-variation-employing-pca-in-scikit-learn-6711e0a5c0b7 +from sklearn.decomposition import PCA +#import tensorflow as tf +#from tensorflow import keras +import numpy as np +import pandas as pd +import seaborn as sns +from sklearn.metrics import matthews_corrcoef + +# pca = PCA().fit(X) +# plt.plot(np.cumsum(pca.explained_variance_ratio_)) +# plt.xlabel(‘number of components’) +# plt.ylabel(‘cumulative explained variance’) + +# from old scripts +fooD = combined_DF_OS(combined_df) + +numerical_ix = fooD['X'].select_dtypes(include=['int64', 'float64']).columns +numerical_ix +num_featuresL = list(numerical_ix) +numerical_colind = fooD['X'].columns.get_indexer(list(numerical_ix) ) +numerical_colind + +numF = fooD['X'][numerical_ix] + +categorical_ix = fooD['X'].select_dtypes(include=['object', 'bool']).columns +categorical_ix +categorical_colind = fooD['X'].columns.get_indexer(list(categorical_ix)) +categorical_colind + +############## + +X_train,X_test,y_train,y_test=train_test_split(numF,fooD['y'],test_size=0.2) + +pca=PCA(n_components=50) +X_train_new=pca.fit_transform(X_train) +X_test_new=pca.transform(X_test) +print(X_train.shape) +print(X_train_new.shape) + +pca.explained_variance_ratio_ +clf=KNeighborsClassifier(n_neighbors=5) +clf.fit(X_train_new,y_train) +y_pred_new=clf.predict(X_test_new) +matthews_corrcoef(y_test,y_pred_new) diff --git a/scripts/ml/untitled6.py b/scripts/ml/untitled6.py new file mode 100644 index 0000000..a6933d4 --- /dev/null +++ b/scripts/ml/untitled6.py @@ -0,0 +1,136 @@ +#!/usr/bin/env python3 +# -*- coding: utf-8 -*- +""" +Created on Fri Sep 2 11:30:18 2022 + +@author: tanu +""" +#https://github.com/yuneeham/PCA-and-feature-selection_sklearn/blob/main/Report%20-%20PCA%20and%20Feature%20Selection.pdf + +#Load Libraries +import numpy as np +import pandas as pd +from sklearn.decomposition import PCA +from sklearn import datasets +from sklearn.preprocessing import scale +from sklearn.preprocessing import StandardScaler +from sklearn.model_selection import train_test_split +from sklearn.ensemble import RandomForestRegressor +from sklearn import metrics +from sklearn.ensemble import RandomForestClassifier +from sklearn.metrics import classification_report +import matplotlib.pyplot as plt +from sklearn.feature_selection import SelectFromModel + +#Load Data +df = pd.read_csv("/home/tanu/Downloads/data.csv") +X = df.loc[:, ' ROA(C) before interest and depreciation before interest':' Equity to Liability'].values +y = df.loc[:,['Bankrupt?']].values + +fn = df.loc[:, ' ROA(C) before interest and depreciation before interest':' Equity to Liability'].keys() + +#Scaler/normalize +scaler = StandardScaler() +Xn = scaler.fit_transform(X) + +#PCA +pca_prep = PCA().fit(Xn) +pca_prep.n_components_ + + +#PCA Explained Variance +pca_prep.explained_variance_ +plt.plot(pca_prep.explained_variance_ratio_) + +#Graph plot - PCA components +plt.plot(pca_prep.explained_variance_ratio_) +plt.xlabel('k number of components') +plt.ylabel('Explained variance') +plt.grid(True) +plt.show() + +#Number of components +n_pc = 17 +pca = PCA(n_components = n_pc).fit(Xn) +Xp = pca.transform(Xn) +print(f'After PCA, we use {pca.n_components_} components. \n') + + +# Split the data into training and testing subsets. +X_train, X_test, y_train, y_test = train_test_split(X,y,test_size =.2,random_state=1234,stratify=y) +Xp_train, Xp_test, yp_train, yp_test = train_test_split(Xp,y,test_size =.2,random_state=1234,stratify=y) + + +#Random Forest Model +rfcm = RandomForestClassifier().fit(X_train, y_train) #Original Data +rfcm_p = RandomForestClassifier().fit(Xp_train, yp_train) #Reduced Data + + +#Prediction +y_pred = rfcm.predict(X_test) +y_pred_p = rfcm_p.predict(Xp_test) + + +# Compare the performance of each mode +report_original = classification_report(y_test, y_pred) +report_pca = classification_report(yp_test, y_pred_p) +print(f'Classification Report - original\n{report_original}') +print(f'Classification Report - pca\n{report_pca}') + + + +## Feature selection and performance comparison + +# Draw a bar chart to see the sorted importance values with feature names. +# Horizontal Bar Chart +# %matplotlib auto +# %matplotlib inline + +importances = rfcm.feature_importances_ +np.sum(importances) +plt.barh(fn,importances) + +df_importances = pd.DataFrame(data=importances, index=fn, + columns=['importance_value']) +df_importances.sort_values(by = 'importance_value', ascending=True, + inplace=True) + +plt.barh(df_importances.index,df_importances.importance_value) + + +# Build a model with a subset of those features. +selector = SelectFromModel(estimator=RandomForestClassifier(),threshold=0.015) +X_reduced = selector.fit_transform(X,y) +selector.threshold_ +selected_TF = selector.get_support() +print(f'\n** {selected_TF.sum()} features are selected.') +X_reduced.shape + + +# Show those selected features. +selected_features = [] +for i,j in zip(selected_TF, fn): + if i: selected_features.append(j) +print(f'Selected Features: {selected_features}') + + +# First 5 features +print(selected_features[0:5]) + + +# Build a model using those reduced number of features. +X_reduced_train, X_reduced_test, y_reduced_train, y_reduced_test \ + = train_test_split(X_reduced,y,test_size =.3, stratify=y) + + +# Build a model with the reduced number of features. +rfcm2 = RandomForestClassifier().fit(X_reduced_train, y_reduced_train) +y_reduced_pred = rfcm2.predict(X_reduced_test) + + +#Classification for Reduced Data +print('\nClassification Report after feature reduction\n') +print(metrics.classification_report(y_reduced_test,y_reduced_pred)) + + + diff --git a/scripts/ml/untitled7_boruta.py b/scripts/ml/untitled7_boruta.py new file mode 100644 index 0000000..c315ee0 --- /dev/null +++ b/scripts/ml/untitled7_boruta.py @@ -0,0 +1,221 @@ +#!/usr/bin/env python3 +# -*- coding: utf-8 -*- +""" +Created on Fri Sep 2 12:13:53 2022 + +@author: tanu +""" +# https://analyticsindiamag.com/hands-on-guide-to-automated-feature-selection-using-boruta/ +import pandas as pd +import numpy as np +from sklearn.ensemble import RandomForestClassifier +from boruta import BorutaPy +from sklearn.model_selection import train_test_split +from sklearn.metrics import accuracy_score +from sklearn.metrics import matthews_corrcoef + + +URL = "https://raw.githubusercontent.com/Aditya1001001/English-Premier-League/master/pos_modelling_data.csv" +data = pd.read_csv(URL) +data.info() +X = data.drop('Position', axis = 1) +y = data['Position'] +X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = .2, random_state = 1) + +rf_all_features = RandomForestClassifier(random_state=1, n_estimators=1000, max_depth=5) +rf_all_features.fit(X_train, y_train) + + +y_pred = rf_all_features.predict(X_test) + +accuracy_score(y_test, rf_all_features.predict(X_test)) +accuracy_score(y_test, y_pred) +matthews_corrcoef(y_test, rf_all_features.predict(X_test)) + +# BORUTA +rfc = RandomForestClassifier(random_state=1, n_estimators=1000, max_depth=5) +boruta_selector = BorutaPy(rfc, n_estimators='auto', verbose=2, random_state=1) +boruta_selector.fit(np.array(X_train), np.array(y_train)) + +# Tells you how many features: GOOD +print("Ranking: ",boruta_selector.ranking_) +print("No. of significant features: ", boruta_selector.n_features_) + +selected_rf_features = pd.DataFrame({'Feature':list(X_train.columns), + 'Ranking':boruta_selector.ranking_}) + +# tells you the ranking: GOOD +selected_rf_features.sort_values(by='Ranking') + +X_important_train = boruta_selector.transform(np.array(X_train)) +X_important_test = boruta_selector.transform(np.array(X_test)) + +rf_boruta = RandomForestClassifier(random_state=1, n_estimators=1000, max_depth=5) +rf_boruta.fit(X_important_train, y_train) +accuracy_score(y_test, rf_boruta.predict(X_important_test)) +matthews_corrcoef(y_test, rf_boruta.predict(X_important_test)) + + + +############################################################################## +# my data : ONLY numerical values +# from old scripts (cm_logo_skf_v2.py) +fooD = combined_DF_OS(combined_df) + +allF = fooD['X'] +numerical_ix = fooD['X'].select_dtypes(include=['int64', 'float64']).columns +numerical_ix +# just numerical for X_train and X_test +X_train_numF = fooD['X'][numerical_ix] +X_test_numF = fooD['X_bts'][numerical_ix] +#X_train = allF + +X_train = X_train_numF +X_test = X_test_numF + +y_train = fooD['y'] +y_test = fooD['y_bts'] + +# 1 model +rf_all_features = RandomForestClassifier(random_state=1, n_estimators=1000, max_depth=5) +rf_all_features.fit(X_train, y_train) + +accuracy_score(y_test, rf_all_features.predict(X_test)) +matthews_corrcoef(y_test, rf_all_features.predict(X_test)) + + +# BORUTA +rfc = RandomForestClassifier(random_state=1, n_estimators=1000, max_depth=5) +boruta_selector = BorutaPy(rfc, n_estimators='auto', verbose=2, random_state=1) +boruta_selector.fit(np.array(X_train), np.array(y_train)) + +# Tells you how many features: GOOD +print("Ranking: ",boruta_selector.ranking_) +print("No. of significant features: ", boruta_selector.n_features_) + +selected_rf_features = pd.DataFrame({'Feature':list(X_train.columns), + 'Ranking':boruta_selector.ranking_}) + +# tells you the ranking: GOOD +selected_rf_features.sort_values(by='Ranking') + +X_important_train = boruta_selector.transform(np.array(X_train)) +X_important_test = boruta_selector.transform(np.array(X_test)) + +rf_boruta = RandomForestClassifier(random_state=1, n_estimators=1000, max_depth=5) +rf_boruta.fit(X_important_train, y_train) +accuracy_score(y_test, rf_boruta.predict(X_important_test)) +matthews_corrcoef(y_test, rf_boruta.predict(X_important_test)) + +############################################################################## +# my data : using both numerical and categorical +# from old scripts (cm_logo_skf_v2.py) +fooD = combined_DF_OS(combined_df) + +numerical_ix = fooD['X'].select_dtypes(include=['int64', 'float64']).columns +numerical_ix +print("\nNo. of numerical indices:", len(numerical_ix)) + +categorical_ix = fooD['X'].select_dtypes(include=['object', 'bool']).columns +categorical_ix +print("\nNo. of categorical indices:", len(categorical_ix)) + + +var_type = "mixeds" + +if var_type == 'mixed': + + t = [('num', MinMaxScaler(), numerical_ix) + , ('cat', OneHotEncoder(), categorical_ix)] + +col_transform = ColumnTransformer(transformers = t + , remainder='passthrough') +#--------------ALEX help +# col_transform +# col_transform.fit(X) +# test = col_transform.transform(X) +# print(col_transform.get_feature_names_out()) + +# foo = col_transform.fit_transform(X) +Xm_train = col_transform.fit_transform(fooD['X']) +fooD['X'].shape +Xm_train.shape + +Xm_test = col_transform.fit_transform(fooD['X_bts']) +fooD['X_bts'].shape +Xm_test.shape + +X_train = Xm_train.copy() +X_test = Xm_test.copy() +X_train.shape +X_test.shape + +y_train = fooD['y'] +y_test = fooD['y_bts'] +y_train.shape +y_test.shape + +# perhaps +#col_transform.fit(fooD['X']) +#encoded_colnames = pd.Index(col_transform.get_feature_names_out()) +#====================== +# 1 model +rf_all_features = RandomForestClassifier(random_state=1, n_estimators=1000, max_depth=5) +rf_all_features.fit(X_train, y_train) + +accuracy_score(y_test, rf_all_features.predict(X_test)) +matthews_corrcoef(y_test, rf_all_features.predict(X_test)) + + +# BORUTA +rfc = RandomForestClassifier(random_state=1, n_estimators=1000, max_depth=5) +boruta_selector = BorutaPy(rfc, n_estimators='auto', verbose=2, random_state=1) +boruta_selector.fit(np.array(X_train), np.array(y_train)) + +# Tells you how many features: GOOD +print("Ranking: ",boruta_selector.ranking_) +print("No. of significant features: ", boruta_selector.n_features_) + +#selected_rf_features = pd.DataFrame({'Feature':list(X_train.columns), +# 'Ranking':boruta_selector.ranking_}) + +# tells you the ranking: GOOD +foo2 = selected_rf_features.sort_values(by='Ranking') + +X_important_train = boruta_selector.transform(np.array(X_train)) +X_important_test = boruta_selector.transform(np.array(X_test)) + +rf_boruta = RandomForestClassifier(random_state=1, n_estimators=1000, max_depth=5) +rf_boruta.fit(X_important_train, y_train) +accuracy_score(y_test, rf_boruta.predict(X_important_test)) +matthews_corrcoef(y_test, rf_boruta.predict(X_important_test)) +################################## +# trying to one hot encode at start +# perhaps +#col_transform.fit(fooD['X']) +#encoded_colnames = pd.Index(col_transform.get_feature_names_out()) + +# def encode_and_bind(original_dataframe, feature_to_encode): +# dummies = pd.get_dummies(original_dataframe[[feature_to_encode]]) +# res = pd.concat([original_dataframe, dummies], axis=1) +# res = res.drop([feature_to_encode], axis=1) +# return(res) + +# features_to_encode = ['feature_1', 'feature_2', 'feature_3', +# 'feature_4'] + +# features_to_encode = list(categorical_ix.copy()) + +# for feature in features_to_encode: +# X_train_enc = encode_and_bind(fooD['X'], feature) +# X_test_enc = encode_and_bind(fooD['X_bts'], feature) + +# c1 = X_train_enc.columns +# c2 = X_test_enc.columns +# X_train_enc.shape +# X_test_enc.shape + +# This one is better! +a = pd.get_dummies(combined_df, columns=features_to_encode) +a1=a.columns +a2 = a.drop(['gene_name', 'dst', 'dst_mode'])