tanu/LSHTM_analysis
1
0
Fork 0
Code Issues Pull requests Releases Wiki Activity

added option to add confusion matrix and target numbers in the mult function

Browse source
This commit is contained in:
Tanushree Tunstall 2022-06-20 17:08:22 +01:00
parent 905327bf4e
commit 135efcee41
3 changed files with 144 additions and 140 deletions
Download patch file Download diff file Expand all files Collapse all files

109
scripts/ml/MultModelsCl.py
View file

@ -138,94 +138,75 @@ def MultModelsCl(input_df, target, skf_cv
, remainder='passthrough') , remainder='passthrough')
# Specify multiple Classification models # Specify multiple Classification models
lr = LogisticRegression(**rs) models = [('Logistic Regression' , LogisticRegression(**rs) )
lrcv = LogisticRegressionCV(**rs) , ('Logistic RegressionCV' , LogisticRegressionCV(**rs) )
gnb = GaussianNB() , ('Gaussian NB' , GaussianNB() )
nb = BernoulliNB() , ('Naive Bayes' , BernoulliNB() )
knn = KNeighborsClassifier() , ('K-Nearest Neighbors' , KNeighborsClassifier() )
svc = SVC(**rs) , ('SVC' , SVC(**rs) )
mlp = MLPClassifier(max_iter = 500, **rs) , ('MLP' , MLPClassifier(max_iter = 500, **rs) )
dt = DecisionTreeClassifier(**rs) , ('Decision Tree' , DecisionTreeClassifier(**rs) )
ets = ExtraTreesClassifier(**rs) , ('Extra Trees' , ExtraTreesClassifier(**rs) )
et = ExtraTreeClassifier(**rs) , ('Extra Tree' , ExtraTreeClassifier(**rs) )
rf = RandomForestClassifier(**rs, n_estimators = 1000 ) , ('Random Forest' , RandomForestClassifier(**rs, n_estimators = 1000 ) )
rf2 = RandomForestClassifier( , ('Random Forest2' , RandomForestClassifier(min_samples_leaf = 5
min_samples_leaf = 5
, n_estimators = 1000 , n_estimators = 1000
, bootstrap = True , bootstrap = True
, oob_score = True , oob_score = True
, **njobs , **njobs
, **rs , **rs
, max_features = 'auto') , max_features = 'auto') )
xgb = XGBClassifier(**rs, verbosity = 0, use_label_encoder =False) , ('XGBoost' , XGBClassifier(**rs, verbosity = 0, use_label_encoder =False) )
, ('LDA' , LinearDiscriminantAnalysis() )
lda = LinearDiscriminantAnalysis() , ('Multinomial' , MultinomialNB() )
, ('Passive Aggresive' , PassiveAggressiveClassifier(**rs, **njobs) )
mnb = MultinomialNB() , ('Stochastic GDescent' , SGDClassifier(**rs, **njobs) )
, ('AdaBoost Classifier' , AdaBoostClassifier(**rs) )
pa = PassiveAggressiveClassifier(**rs, **njobs) , ('Bagging Classifier' , BaggingClassifier(**rs, **njobs, bootstrap = True, oob_score = True) )
, ('Gaussian Process' , GaussianProcessClassifier(**rs) )
sgd = SGDClassifier(**rs, **njobs) , ('Gradient Boosting' , GradientBoostingClassifier(**rs) )
, ('QDA' , QuadraticDiscriminantAnalysis() )
abc = AdaBoostClassifier(**rs) , ('Ridge Classifier' , RidgeClassifier(**rs) )
bc = BaggingClassifier(**rs, **njobs, bootstrap = True, oob_score = True) , ('Ridge ClassifierCV' , RidgeClassifierCV(cv = 10) )
gpc = GaussianProcessClassifier(**rs)
gbc = GradientBoostingClassifier(**rs)
qda = QuadraticDiscriminantAnalysis()
rc = RidgeClassifier(**rs)
rccv = RidgeClassifierCV(cv = 10)
models = [('Logistic Regression' , lr)
, ('Logistic RegressionCV' , lrcv)
, ('Gaussian NB' , gnb)
, ('Naive Bayes' , nb)
, ('K-Nearest Neighbors' , knn)
, ('SVC' , svc)
, ('MLP' , mlp)
, ('Decision Tree' , dt)
, ('Extra Trees' , ets)
, ('Extra Tree' , et)
, ('Random Forest' , rf)
, ('Random Forest2' , rf2)
, ('XGBoost' , xgb)
, ('LDA' , lda)
, ('Multinomial' , mnb)
, ('Passive Aggresive' , pa)
, ('Stochastic GDescent' , sgd)
, ('AdaBoost Classifier' , abc)
, ('Bagging Classifier' , bc)
, ('Gaussian Process' , gpc)
, ('Gradient Boosting' , gbc)
, ('QDA' , qda)
, ('Ridge Classifier' , rc)
, ('Ridge ClassifierCV' , rccv)
] ]
mm_skf_scoresD = {} mm_skf_scoresD = {}
print('\n==============================================================\n'
, '\nRunning several classification models (n):', len(models)
,'\nList of models:')
for m in models:
print(m)
print('\n================================================================\n')
index = 1
for model_name, model_fn in models: for model_name, model_fn in models:
print('\nModel_name:', model_name print('\nRunning classifier:', index
, '\nModel func:' , model_fn , '\nModel_name:' , model_name
, '\nList of models:', models) , '\nModel func:' , model_fn)
index = index+1
model_pipeline = Pipeline([ model_pipeline = Pipeline([
('prep' , col_transform) ('prep' , col_transform)
, ('model' , model_fn)]) , ('model' , model_fn)])
print('Running model pipeline:', model_pipeline) print('\nRunning model pipeline:', model_pipeline)
skf_cv_mod = cross_validate(model_pipeline skf_cv_modD = cross_validate(model_pipeline
, input_df , input_df
, target , target
, cv = skf_cv , cv = skf_cv
, scoring = scoring_fn , scoring = scoring_fn
, return_train_score = True) , return_train_score = True)
#==============================
# Extract mean values for CV
#==============================
mm_skf_scoresD[model_name] = {} mm_skf_scoresD[model_name] = {}
for key, value in skf_cv_mod.items():
for key, value in skf_cv_modD.items():
print('\nkey:', key, '\nvalue:', value) print('\nkey:', key, '\nvalue:', value)
print('\nmean value:', mean(value)) print('\nmean value:', mean(value))
mm_skf_scoresD[model_name][key] = round(mean(value),2) mm_skf_scoresD[model_name][key] = round(mean(value),2)
#pp.pprint(mm_skf_scoresD)
#cvtrain_mcc = mm_skf_scoresD[model_name]['test_mcc']
#return(mm_skf_scoresD) #return(mm_skf_scoresD)
#%% #%%

101
scripts/ml/MultModelsCl_dissected.py
View file

@ -101,6 +101,9 @@ jacc_score_fn = {'jcc': make_scorer(jaccard_score)}
def MultModelsCl_dissected(input_df, target, skf_cv def MultModelsCl_dissected(input_df, target, skf_cv
, blind_test_input_df , blind_test_input_df
, blind_test_target , blind_test_target
, add_cm = True # adds confusion matrix based on cross_val_predict
, add_yn = True # adds target var class numbers
, feature_groups = ['']
, var_type = ['numerical', 'categorical','mixed']): , var_type = ['numerical', 'categorical','mixed']):
''' '''
@ -201,52 +204,88 @@ def MultModelsCl_dissected(input_df, target, skf_cv
, scoring = scoring_fn , scoring = scoring_fn
, return_train_score = True) , return_train_score = True)
#---------- #######################################################################
# check 1 #======================================================
#---------- # Option 1: Add confusion matrix from cross_val_predict
foo_df = pd.DataFrame.from_dict(skf_cv_modD, orient ='index') # Understand and USE with caution
#foo_df = pd.DataFrame.from_dict(skf_cv_modD)
#===================
# Confusion matrix: Not an easy problem to solve! STILL DOING it, USE with caution
# cross_val_score, cross_val_predict, "Passing these predictions into an evaluation metric may not be a valid way to measure generalization performance. Results can differ from cross_validate and cross_val_score unless all tests sets have equal size and the metric decomposes over samples." # cross_val_score, cross_val_predict, "Passing these predictions into an evaluation metric may not be a valid way to measure generalization performance. Results can differ from cross_validate and cross_val_score unless all tests sets have equal size and the metric decomposes over samples."
# https://stackoverflow.com/questions/65645125/producing-a-confusion-matrix-with-cross-validate # https://stackoverflow.com/questions/65645125/producing-a-confusion-matrix-with-cross-validate
#=================== #======================================================
y_pred = cross_val_predict(model_pipeline, input_df, target, cv = 10, **njobs) if add_cm:
#-----------------------------------------------------------
# Initialise dict of Confusion Matrix (cm)
#-----------------------------------------------------------
cmD = {}
# Calculate cm
y_pred = cross_val_predict(model_pipeline, input_df, target, cv = skf_cv, **njobs)
#_tn, _fp, _fn, _tp = confusion_matrix(y_pred, y).ravel() # internally #_tn, _fp, _fn, _tp = confusion_matrix(y_pred, y).ravel() # internally
tn, fp, fn, tp = confusion_matrix(y_pred, target).ravel() tn, fp, fn, tp = confusion_matrix(y_pred, target).ravel()
# create a dict of confusion matrix that can be appended to the one above
# cmD = {'TN' : np.array(tn) # Build dict
# , 'FP': np.array(fp)
# , 'FN': np.array(fn)
# , 'TP': np.array(tp)}
cmD = {'TN' : tn cmD = {'TN' : tn
, 'FP': fp , 'FP': fp
, 'FN': fn , 'FN': fn
, 'TP': tp} , 'TP': tp}
#---------------------------------
# Update cv dict with cmD and tbtD
#----------------------------------
skf_cv_modD.update(cmD) skf_cv_modD.update(cmD)
else:
skf_cv_modD = skf_cv_modD
#######################################################################
#=============================================
# Option 2: Add targety numbers for data
#=============================================
if add_yn:
#---------- #-----------------------------------------------------------
# check 2 # Initialise dict of target numbers: training and blind (tbt)
#---------- #-----------------------------------------------------------
#foo2_df = pd.DataFrame.from_dict(skf_cv_modD, orient ='index') tbtD = {}
#foo_df = pd.DataFrame.from_dict(skf_cv_modD)
# training y
tyn = Counter(target)
tyn_neg = tyn[0]
tyn_pos = tyn[1]
# blind test y
btyn = Counter(blind_test_target)
btyn_neg = btyn[0]
btyn_pos = btyn[1]
# Build dict
tbtD = {'trainingY_neg' : tyn_neg
, 'trainingY_pos' : tyn_pos
, 'blindY_neg' : btyn_neg
, 'blindY_pos' : btyn_pos}
#---------------------------------
# Update cv dict with cmD and tbtD
#----------------------------------
skf_cv_modD.update(tbtD)
else:
skf_cv_modD = skf_cv_modD
#######################################################################
#==============================
# Extract mean values for CV
#==============================
mm_skf_scoresD[model_name] = {} mm_skf_scoresD[model_name] = {}
for key, value in skf_cv_modD.items(): for key, value in skf_cv_modD.items():
print('\nkey:', key, '\nvalue:', value) print('\nkey:', key, '\nvalue:', value)
print('\nmean value:', np.mean(value)) print('\nmean value:', np.mean(value))
mm_skf_scoresD[model_name][key] = round(np.mean(value),2) mm_skf_scoresD[model_name][key] = round(np.mean(value),2)
#return(mm_skf_scoresD) #return(mm_skf_scoresD)
#%% #%%
#========================= #=========================
# Blind test: BTS results # Blind test: BTS results
#========================= #=========================
# Build the final results with all scores for a feature selected model # Build the final results with all scores for the model
#bts_predict = gscv_fs.predict(blind_test_input_df) #bts_predict = gscv_fs.predict(blind_test_input_df)
model_pipeline.fit(input_df, target) model_pipeline.fit(input_df, target)
bts_predict = model_pipeline.predict(blind_test_input_df) bts_predict = model_pipeline.predict(blind_test_input_df)
@ -255,22 +294,6 @@ def MultModelsCl_dissected(input_df, target, skf_cv
print('\nMCC on Blind test:' , bts_mcc_score) print('\nMCC on Blind test:' , bts_mcc_score)
print('\nAccuracy on Blind test:', round(accuracy_score(blind_test_target, bts_predict),2)) print('\nAccuracy on Blind test:', round(accuracy_score(blind_test_target, bts_predict),2))
# Diff b/w train and bts test scores
#train_test_diff_MCC = cvtrain_mcc - bts_mcc_score
# print('\nDiff b/w train and blind test score (MCC):', train_test_diff)
# # create a dict with all scores
# lr_btsD = { 'model_name': model_name
# , 'bts_mcc':None
# , 'bts_fscore':None
# , 'bts_precision':None
# , 'bts_recall':None
# , 'bts_accuracy':None
# , 'bts_roc_auc':None
# , 'bts_jaccard':None}
mm_skf_scoresD[model_name]['bts_mcc'] = bts_mcc_score mm_skf_scoresD[model_name]['bts_mcc'] = bts_mcc_score
mm_skf_scoresD[model_name]['bts_fscore'] = round(f1_score(blind_test_target, bts_predict),2) mm_skf_scoresD[model_name]['bts_fscore'] = round(f1_score(blind_test_target, bts_predict),2)
mm_skf_scoresD[model_name]['bts_precision'] = round(precision_score(blind_test_target, bts_predict),2) mm_skf_scoresD[model_name]['bts_precision'] = round(precision_score(blind_test_target, bts_predict),2)

40
scripts/ml/pnca_config_dissected.py
View file

@ -104,29 +104,29 @@ else:
print('\n#####################################################################\n') print('\n#####################################################################\n')
############################################################################### # ###############################################################################
#================== # #==================
# Baseline models # # Baseline models
#================== # #==================
mm_skf_scoresD = MultModelsCl(input_df = X # mm_skf_scoresD = MultModelsCl(input_df = X
, target = y # , target = y
, var_type = 'mixed' # , var_type = 'mixed'
, skf_cv = skf_cv # , skf_cv = skf_cv
, blind_test_input_df = X_bts # , blind_test_input_df = X_bts
, blind_test_target = y_bts) # , blind_test_target = y_bts)
baseline_all = pd.DataFrame(mm_skf_scoresD) # baseline_all = pd.DataFrame(mm_skf_scoresD)
baseline_all = baseline_all.T # baseline_all = baseline_all.T
#baseline_train = baseline_all.filter(like='train_', axis=1) # #baseline_train = baseline_all.filter(like='train_', axis=1)
baseline_CT = baseline_all.filter(like='test_', axis=1) # baseline_CT = baseline_all.filter(like='test_', axis=1)
baseline_CT.sort_values(by=['test_mcc'], ascending=False, inplace=True) # baseline_CT.sort_values(by=['test_mcc'], ascending=False, inplace=True)
baseline_BT = baseline_all.filter(like='bts_', axis=1) # baseline_BT = baseline_all.filter(like='bts_', axis=1)
baseline_BT.sort_values(by = ['bts_mcc'], ascending = False, inplace = True) # baseline_BT.sort_values(by = ['bts_mcc'], ascending = False, inplace = True)
# Write csv # # Write csv
baseline_CT.to_csv(outdir_ml + gene.lower() + '_baseline_CT_allF.csv') # baseline_CT.to_csv(outdir_ml + gene.lower() + '_baseline_CT_allF.csv')
baseline_BT.to_csv(outdir_ml + gene.lower() + '_baseline_BT_allF.csv') # baseline_BT.to_csv(outdir_ml + gene.lower() + '_baseline_BT_allF.csv')
# #%% SMOTE NC: Oversampling [Numerical + categorical] # #%% SMOTE NC: Oversampling [Numerical + categorical]