IE6400_Quiz22_Day24

Building and Validating a Predictive Maintenance Model Data Loading Load the training and validation datasets for analysis. import pandas as pd # Load the datasets training_data = pd . read_csv ( 'training_dataset.csv' ) validation_data = pd . read_csv ( 'validation_dataset.csv' ) # Display the first few rows of the training dataset print ( training_data . head ()) # Display the first few rows of the validation dataset print ( validation_data . head ()) Feature_1 Feature_2 Feature_3 Feature_4 Feature_5 Feature_6 \ 0 0.155259 0.660369 2.175290 6.258942 16.098589 16.099171 1 0.610906 0.992507 2.672643 7.621450 18.223143 13.329514 2 1.183112 0.679466 2.968680 9.809290 19.919326 9.043429 3 0.614589 0.647379 1.974244 5.210570 15.222118 16.993711 4 0.404007 0.661153 2.552086 7.244011 18.242283 13.341818 Feature_7 Feature_8 Feature_9 Feature_10 ... Feature_492 Feature_493 \ 0 1.470314 -4.734506 -7.213391 -8.575321 ... -0.507059 -2.372885 1 -0.799513 -5.592666 -7.301702 -8.757725 ... 5.088316 -1.256212 2 -2.500281 -6.075475 -7.412497 -8.905431 ... -8.930852 -13.528654 3 2.509928 -4.549607 -6.797826 -8.202687 ... -12.580317 -10.396995 4 -0.544337 -5.363483 -7.508647 -8.837681 ... -0.790838 -1.699379 Feature_494 Feature_495 Feature_496 Feature_497 Feature_498 \ 0 -5.036329 -12.242622 -16.960816 -6.879618 0.922027 1 -4.228322 -6.540655 -9.827748 -12.505886 -9.542722 2 -11.426239 -5.573573 -2.059390 -1.787930 -2.806238 3 -5.641676 -3.976051 -4.147271 -5.824191 -10.471274 4 -3.411248 -6.849308 -12.954395 -14.061112 -6.753582 Feature_499 Feature_500 Status 0 3.393410 6.081871 1.0 1 -5.376019 -3.338145 1.0 2 -5.660278 -11.730715 1.0 3 -13.132935 -9.758947 1.0 4 -1.821272 -0.415765 1.0 [5 rows x 501 columns] Feature_1 Feature_2 Feature_3 Feature_4 Feature_5 Feature_6 \ 0 0.123080 0.860825 3.173313 9.917909 19.849869 9.099698 1 0.750959 1.619463 5.148140 13.948681 18.074979 3.749289 2 0.698000 1.473921 4.351274 12.652395 19.132758 5.508475 3 0.297695 0.549153 1.835041 5.975847 16.434054 16.217664 4 0.739391 0.763177 2.644591 8.152586 19.094034 11.789167 Feature_7 Feature_8 Feature_9 Feature_10 ... Feature_492 Feature_493 \ 0 -2.314409 -6.131037 -7.831784 -9.115000 ... -12.035100 -16.167343 1 -3.971191 -6.643023 -8.402794 -9.505138 ... 2.672961 5.306708 2 -3.634955 -6.763107 -8.170255 -9.328350 ... -3.722906 -3.587257 3 1.527474 -4.933753 -7.267388 -8.455417 ... 2.895946 7.284208 4 -1.122723 -5.634264 -7.380879 -8.873741 ... -3.309775 -4.810450 Feature_494 Feature_495 Feature_496 Feature_497 Feature_498 \ 0 -7.498560 -0.606933 1.874816 3.757005 6.917021 1 10.058428 14.127538 10.126872 4.183676 1.602770 2 -5.185838 -8.922863 -13.068522 -11.212529 -5.625227 3 14.978634 14.149709 3.603407 -1.436238 -3.579593 4 -9.232337 -13.327797 -11.236511 -5.419281 -2.618585 Feature_499 Feature_500 Status 0 11.778411 13.535583 1.0 1 1.625422 2.527926 1.0 2 -2.847045 -2.721370 1.0 3 -6.033314 -10.311425 1.0 4 -2.355058 -3.467571 1.0 [5 rows x 501 columns] Data Exploration and Preprocessing Conduct an exploration of the datasets to understand their structure. Prepare the data for modeling: normalization, handling missing values, etc. import pandas as pd import numpy as np from sklearn.preprocessing import StandardScaler from sklearn.model_selection import train_test_split from sklearn.metrics import accuracy_score , precision_score , recall_score , f1_score from sklearn.neural_network import MLPClassifier from sklearn.ensemble import RandomForestClassifier from sklearn.svm import SVC import matplotlib.pyplot as plt import seaborn as sns # Explore the structure of the datasets print ( training_data . info ()) print ( validation_data . info ()) <class 'pandas.core.frame.DataFrame'> RangeIndex: 300 entries, 0 to 299 Columns: 501 entries, Feature_1 to Status dtypes: float64(501) memory usage: 1.1 MB None <class 'pandas.core.frame.DataFrame'> RangeIndex: 300 entries, 0 to 299 Columns: 501 entries, Feature_1 to Status dtypes: float64(501) memory usage: 1.1 MB None # Check for missing values print ( training_data . isnull () . sum ()) print ( validation_data . isnull () . sum ()) Feature_1 0 Feature_2 0 Feature_3 0 Feature_4 0 Feature_5 0 .. Feature_497 0 Feature_498 0 Feature_499 0 Feature_500 0 Status 0 Length: 501, dtype: int64 Feature_1 0 Feature_2 0 Feature_3 0 Feature_4 0 Feature_5 0 .. Feature_497 0 Feature_498 0 Feature_499 0 Feature_500 0 Status 0 Length: 501, dtype: int64 # Prepare the data for modeling X_train = training_data . drop ( 'Status' , axis = 1 ) y_train = training_data [ 'Status' ] X_val = validation_data . drop ( 'Status' , axis = 1 ) y_val = validation_data [ 'Status' ] # Normalization scaler = StandardScaler () X_train_scaled = scaler . fit_transform ( X_train ) X_val_scaled = scaler . transform ( X_val ) Feature Extraction Method Implement Recurrence Quantification Analysis (RQA) and Network measurements for advanced feature extraction from the training data. import pandas as pd import numpy as np from sklearn.preprocessing import StandardScaler from sklearn.model_selection import train_test_split from sklearn.neural_network import MLPClassifier from sklearn.metrics import accuracy_score , precision_score , recall_score , f1_score import networkx as nx from pyrqa.time_series import SingleTimeSeries from pyrqa.settings import Settings from pyrqa.analysis_type import Classic # Load the datasets training_data = pd . read_csv ( 'training_dataset.csv' ) validation_data = pd . read_csv ( 'validation_dataset.csv' ) # Prepare the data for modeling X_train = training_data . drop ( 'Status' , axis = 1 ) y_train = training_data [ 'Status' ] X_val = validation_data . drop ( 'Status' , axis = 1 ) y_val = validation_data [ 'Status' ] # Normalization scaler = StandardScaler () X_train_scaled = scaler . fit_transform ( X_train ) X_val_scaled = scaler . transform ( X_val ) # Recurrence Quantification Analysis (RQA) def apply_rqa ( data ): time_series = SingleTimeSeries ( data . values . flatten ()) settings = Settings ( time_series , analysis_type = Classic ( minimal_dynamic_threshold = 0.01 , method = "fan" , neighbourhood = 0.01 , min_diagonal_line_length = 2 , min_vertical_line_length = 2 , min_white_vertical_line_length = 2 ) ) result = settings . compute () return result # Apply RQA to training data rqa_features_train = X_train . apply ( apply_rqa ) # Apply RQA to validation data rqa_features_val = X_val . apply ( apply_rqa ) # Network measurements (using networkx) def calculate_network_measures ( data ): correlation_matrix = np . corrcoef ( data , rowvar = False ) graph = nx . from_numpy_array ( correlation_matrix ) # Add more network measurements based on your requirements measures = { "average_clustering" : nx . average_clustering ( graph ), "average_shortest_path_length" : nx . average_shortest_path_length ( graph ), # Add more measures as needed } return measures # Calculate network measures for training data network_features_train = X_train . apply ( calculate_network_measures ) # Calculate network measures for validation data network_features_val = X_val . apply ( calculate_network_measures ) # Combine RQA and network features X_train_enhanced = pd . concat ([ rqa_features_train , network_features_train ], axis = 1 ) X_val_enhanced = pd . concat ([ rqa_features_val , network_features_val ], axis = 1 ) # Model Development # Model A: Using the original feature set model_A = MLPClassifier ( random_state = 42 ) model_A . fit ( X_train_scaled , y_train ) # Model B: Using the enhanced feature set from advanced feature extraction model_B = MLPClassifier ( random_state = 42 ) model_B . fit ( X_train_enhanced , y_train ) # Model Validation # Validate Model A y_val_pred_A = model_A . predict ( X_val_scaled ) # Validate Model B y_val_pred_B = model_B . predict ( X_val_enhanced ) # Result Analysis and Visualization # Compare the performance of both models metrics_A = [ accuracy_score , precision_score , recall_score , f1_score ] metrics_names = [ 'Accuracy' , 'Precision' , 'Recall' , 'F1 Score' ] results = { 'Model' : [], 'Metric' : [], 'Value' : []} # Model A for metric , name in zip ( metrics_A , metrics_names ): result = metric ( y_val , y_val_pred_A ) results [ 'Model' ] . append ( 'Model A' ) results [ 'Metric' ] . append ( name ) results [ 'Value' ] . append ( result ) # Model B for metric , name in zip ( metrics_A , metrics_names ): result = metric ( y_val , y_val_pred_B ) results [ 'Model' ] . append ( 'Model B' ) results [ 'Metric' ] . append ( name ) results [ 'Value' ] . append ( result ) # Create visualizations for performance metrics results_df = pd . DataFrame ( results ) print ( results_df ) # Visualize the comparison of Model A and Model B plt . figure ( figsize = ( 10 , 6 )) sns . barplot ( x = 'Metric' , y = 'Value' , hue = 'Model' , data = results_df ) plt . title ( 'Comparison of Model Performance' ) plt . show () --------------------------------------------------------------------------- ImportError Traceback (most recent call last) Cell In[6], line 8 6 from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score 7 import networkx as nx ----> 8 from pyrqa.time_series import SingleTimeSeries 9 from pyrqa.settings import Settings 10 from pyrqa.analysis_type import Classic ImportError : cannot import name 'SingleTimeSeries' from 'pyrqa.time_series' (/Users/skyleraliya/opt/anaconda3/lib/python3.9/site-packages/pyrqa/time_series.py) pip show pyRQA Name: PyRQA Version: 8.0.0 Summary: Recurrence analysis in a massively parallel manner using the OpenCL framework. Home-page: Author: Tobias Rawald Author-email: pyrqa@gmx.net License: Apache License 2.0 Location: /Users/skyleraliya/opt/anaconda3/lib/python3.9/site-packages Requires: Mako, numpy, Pillow, pyopencl, scipy Required-by: Note: you may need to restart the kernel to use updated packages. # Detailed exploration of the datasets # Function to explore a dataset def explore_data ( df ): exploratory_data = {} exploratory_data [ 'summary' ] = df . describe () exploratory_data [ 'missing_values' ] = df . isnull () . sum () exploratory_data [ 'data_types' ] = df . dtypes return exploratory_data # Explore training data training_exploration = explore_data ( training_data ) # Explore validation data validation_exploration = explore_data ( validation_data ) # Normalize the features in the training and validation datasets # Only the features should be normalized, not the target variable 'Status' from sklearn.preprocessing import StandardScaler # Initialize the StandardScaler scaler = StandardScaler () # Fit the scaler on the training data and transform both training and validation data # We exclude the 'Status' column during scaling training_features = training_data . drop ( columns = [ 'Status' ]) validation_features = validation_data . drop ( columns = [ 'Status' ]) scaler . fit ( training_features ) # Perform the transformation normalized_training_data = scaler . transform ( training_features ) normalized_validation_data = scaler . transform ( validation_features ) # Replace old values with normalized values, keeping the 'Status' column intact training_data_normalized = pd . DataFrame ( normalized_training_data , columns = training_features . columns ) training_data_normalized [ 'Status' ] = training_data [ 'Status' ] validation_data_normalized = pd . DataFrame ( normalized_validation_data , columns = validation_features . columns ) validation_data_normalized [ 'Status' ] = validation_data [ 'Status' ] ( training_exploration , validation_exploration , training_data_normalized . head (), validation_data_normalized . head ()) ({'summary': Feature_1 Feature_2 Feature_3 Feature_4 Feature_5 Feature_6 \ count 300.000000 300.000000 300.000000 300.000000 300.000000 300.000000 mean 0.133587 0.353081 1.148927 3.317685 5.996621 3.559958 std 0.464645 0.683597 1.762382 4.925212 8.333159 5.400490 min -1.299247 -2.567137 -3.348178 -3.592233 -3.219585 -3.600452 25% -0.119510 -0.027414 0.086244 0.204398 0.364976 0.438351 50% 0.101041 0.249088 0.400961 0.584282 0.687151 0.770921 75% 0.425922 0.838214 2.208981 7.051022 17.101556 5.775406 max 1.437154 1.946377 6.346909 16.082052 20.141440 19.992030 Feature_7 Feature_8 Feature_9 Feature_10 ... Feature_492 \ count 300.000000 300.000000 300.000000 300.000000 ... 300.000000 mean 0.046032 -1.432604 -2.153879 -2.600541 ... -0.401728 std 2.784697 3.404206 3.952983 4.588838 ... 7.185068 min -4.989652 -7.164711 -8.709318 -9.939976 ... -18.490077 25% -1.386086 -5.368501 -7.316010 -8.702373 ... -2.815486 50% 0.473083 0.277924 -0.003062 0.095490 ... -0.825105 75% 0.856203 0.801420 0.864308 0.925883 ... 2.345776 max 19.567093 17.508389 4.250973 4.703203 ... 30.755470 Feature_493 Feature_494 Feature_495 Feature_496 Feature_497 \ count 300.000000 300.000000 300.000000 300.000000 300.000000 mean -0.394811 -0.445478 -0.450620 -0.396666 -0.354362 std 7.224344 7.078942 7.060402 7.041630 7.019486 min -18.238748 -18.861712 -18.839982 -19.969468 -18.730305 25% -2.990391 -3.295071 -3.132898 -2.849080 -3.417853 50% -0.752216 -0.628879 -0.571799 -0.546996 -0.424046 75% 2.197214 2.215800 1.815272 1.702486 1.530649 max 31.672199 32.124138 32.518974 33.409954 33.517280 Feature_498 Feature_499 Feature_500 Status count 300.000000 300.000000 300.000000 300.000000 mean -0.465986 -0.627828 -0.551134 2.000000 std 7.072625 7.298489 7.411132 0.817861 min -18.467430 -19.040604 -19.114384 1.000000 25% -3.537704 -3.896523 -4.135700 1.000000 50% -0.304221 -0.186471 -0.036254 2.000000 75% 1.866116 1.976870 1.988903 3.000000 max 33.039946 33.314707 32.928803 3.000000 [8 rows x 501 columns], 'missing_values': Feature_1 0 Feature_2 0 Feature_3 0 Feature_4 0 Feature_5 0 .. Feature_497 0 Feature_498 0 Feature_499 0 Feature_500 0 Status 0 Length: 501, dtype: int64, 'data_types': Feature_1 float64 Feature_2 float64 Feature_3 float64 Feature_4 float64 Feature_5 float64 ... Feature_497 float64 Feature_498 float64 Feature_499 float64 Feature_500 float64 Status float64 Length: 501, dtype: object}, {'summary': Feature_1 Feature_2 Feature_3 Feature_4 Feature_5 Feature_6 \ count 300.000000 300.000000 300.000000 300.000000 300.000000 300.000000 mean 0.173138 0.423545 1.257116 3.547109 6.052185 3.180508 std 0.405372 0.700782 1.850152 5.261167 8.478668 5.242294 min -1.463762 -2.487939 -1.737156 -2.285522 -3.048860 -2.582134 25% -0.036859 0.071475 0.175340 0.257864 0.415697 0.449917 50% 0.105102 0.269172 0.346894 0.479223 0.580994 0.666464 75% 0.463262 0.861609 2.473455 7.796831 17.104935 4.728514 max 1.128735 2.118145 6.726057 17.136014 19.962632 19.762905 Feature_7 Feature_8 Feature_9 Feature_10 ... Feature_492 \ count 300.000000 300.000000 300.000000 300.000000 ... 300.000000 mean -0.365900 -1.780681 -2.394442 -2.817422 ... -0.790263 std 2.354452 3.154211 3.949272 4.512745 ... 7.807457 min -5.187750 -7.439548 -8.746702 -9.616524 ... -28.605543 25% -1.874546 -5.506959 -7.375137 -8.745140 ... -3.018682 50% 0.493582 -0.078736 -0.261905 -0.316128 ... -0.826995 75% 0.746429 0.768297 0.839751 0.887290 ... 3.014377 max 13.640237 3.371212 3.305640 3.391341 ... 19.788439 Feature_493 Feature_494 Feature_495 Feature_496 Feature_497 \ count 300.000000 300.000000 300.000000 300.000000 300.000000 mean -0.659900 -0.490028 -0.449033 -0.461000 -0.373395 std 7.995240 8.092395 8.208499 8.284201 8.071444 min -28.988463 -29.593378 -30.162245 -30.913387 -31.111843 25% -3.597136 -3.871761 -3.723010 -3.046614 -3.784670 50% -0.746635 -0.685207 -0.575542 -0.466373 -0.374921 75% 3.570288 3.665994 4.371823 4.082067 4.203521 max 19.820185 20.020610 21.382359 21.651101 21.274513 Feature_498 Feature_499 Feature_500 Status count 300.000000 300.000000 300.000000 300.000000 mean -0.336339 -0.580490 -0.790558 2.000000 std 8.010108 8.018548 8.015539 0.817861 min -31.307439 -31.974854 -32.149645 1.000000 25% -3.448527 -3.160960 -3.611293 1.000000 50% -0.257718 -0.143136 -0.019509 2.000000 75% 3.663612 2.905283 2.545813 3.000000 max 21.593234 21.266144 21.462606 3.000000 [8 rows x 501 columns], 'missing_values': Feature_1 0 Feature_2 0 Feature_3 0 Feature_4 0 Feature_5 0 .. Feature_497 0 Feature_498 0 Feature_499 0 Feature_500 0 Status 0 Length: 501, dtype: int64, 'data_types': Feature_1 float64 Feature_2 float64 Feature_3 float64 Feature_4 float64 Feature_5 float64 ... Feature_497 float64 Feature_498 float64 Feature_499 float64 Feature_500 float64 Status float64 Length: 501, dtype: object}, Feature_1 Feature_2 Feature_3 Feature_4 Feature_5 Feature_6 \ 0 0.046721 0.450267 0.583346 0.598182 1.214287 2.325745 1 1.028993 0.936947 0.866022 0.875283 1.469665 1.812035 2 2.262543 0.478249 1.034278 1.320238 1.673551 1.017062 3 1.036933 0.431232 0.469079 0.384968 1.108933 2.491662 4 0.582966 0.451415 0.797502 0.798521 1.471966 1.814317 Feature_7 Feature_8 Feature_9 Feature_10 ... Feature_492 Feature_493 \ 0 0.512322 -0.971569 -1.282061 -1.304200 ... -0.014684 -0.274264 1 -0.304147 -1.224078 -1.304439 -1.344016 ... 0.765367 -0.119435 2 -0.915923 -1.366142 -1.332514 -1.376258 ... -1.189046 -1.821036 3 0.886277 -0.917163 -1.176759 -1.222860 ... -1.697818 -1.386824 4 -0.212359 -1.156642 -1.356878 -1.361469 ... -0.054246 -0.180881 Feature_494 Feature_495 Feature_496 Feature_497 Feature_498 \ 0 -0.649606 -1.672951 -2.356248 -0.931145 0.196579 1 -0.535273 -0.864003 -1.341570 -1.734005 -1.285506 2 -1.553779 -0.726802 -0.236522 -0.204568 -0.331442 3 -0.735262 -0.500159 -0.533523 -0.780537 -1.417013 4 -0.419657 -0.907792 -1.786335 -1.955934 -0.890490 Feature_499 Feature_500 Status 0 0.551889 0.896501 1.0 1 -0.651659 -0.376686 1.0 2 -0.690672 -1.511005 1.0 3 -1.716246 -1.244506 1.0 4 -0.163793 0.018296 1.0 [5 rows x 501 columns], Feature_1 Feature_2 Feature_3 Feature_4 Feature_5 Feature_6 \ 0 -0.022650 0.743993 1.150584 1.342328 1.665202 1.027499 1 1.330917 1.855621 2.273000 2.162092 1.451855 0.035117 2 1.216750 1.642358 1.820092 1.898458 1.579003 0.361406 3 0.353781 0.287302 0.389961 0.540607 1.254611 2.347723 4 1.305979 0.600910 0.850079 0.983304 1.574349 1.526335 Feature_7 Feature_8 Feature_9 Feature_10 ... Feature_492 Feature_493 \ 0 -0.849064 -1.382491 -1.438760 -1.422004 ... -1.621809 -2.186896 1 -1.445017 -1.533140 -1.583451 -1.507165 ... 0.428643 0.790528 2 -1.324071 -1.568474 -1.524527 -1.468575 ... -0.463006 -0.442640 3 0.532883 -1.030196 -1.295744 -1.278027 ... 0.459729 1.064713 4 -0.420408 -1.236318 -1.324502 -1.369341 ... -0.405411 -0.612238 Feature_494 Feature_495 Feature_496 Feature_497 Feature_498 \ 0 -0.998012 -0.022176 0.323118 0.586686 1.045629 1 1.486304 2.068227 1.496972 0.647572 0.292991 2 -0.670761 -1.201971 -1.802570 -1.549445 -0.730685 3 2.182513 2.071373 0.569010 -0.154382 -0.440969 4 -1.243341 -1.826906 -1.541966 -0.722757 -0.304865 Feature_499 Feature_500 Status 0 1.702677 1.903927 1.0 1 0.309244 0.416158 1.0 2 -0.304573 -0.293324 1.0 3 -0.741868 -1.319178 1.0 4 -0.237051 -0.394179 1.0 [5 rows x 501 columns]) from pyrqa.time_series import SingleTimeSeries from pyrqa.settings import Settings from pyrqa.analysis_type import Classic from pyrqa.neighbourhood import FixedRadius from pyrqa.metric import EuclideanMetric from pyrqa.computation import RQAComputation # Example for one feature column time_series = SingleTimeSeries ( training_data [ 'Feature_1' ], embedding_dimension = 2 , time_delay = 1 ) settings = Settings ( time_series , analysis_type = Classic , neighbourhood = FixedRadius ( 0.1 ), similarity_measure = EuclideanMetric , theiler_corrector = 1 ) computation = RQAComputation . create ( settings , verbose = True ) result = computation . run () result . min_diagonal_line_length = 2 result . min_vertical_line_length = 2 result . min_white_vertical_line_length = 2 recurrence_plot = result . recurrence_plot () --------------------------------------------------------------------------- ImportError Traceback (most recent call last) Cell In[9], line 1 ----> 1 from pyrqa.time_series import SingleTimeSeries 2 from pyrqa.settings import Settings 3 from pyrqa.analysis_type import Classic ImportError : cannot import name 'SingleTimeSeries' from 'pyrqa.time_series' (/Users/skyleraliya/opt/anaconda3/lib/python3.9/site-packages/pyrqa/time_series.py) import networkx as nx # Compute the correlation matrix correlation_matrix = training_data . corr () # Create a graph from the correlation matrix threshold = 0.8 # This is an arbitrary threshold for demonstration purposes graph = nx . from_pandas_adjacency ( correlation_matrix [ correlation_matrix > threshold ]) # Compute degree centrality degree_centrality = nx . degree_centrality ( graph ) from sklearn.preprocessing import StandardScaler # Initialize the StandardScaler scaler = StandardScaler () # Fit the scaler on the training data and transform both training and validation data # We exclude the 'Status' column during scaling training_features = training_data . drop ( columns = [ 'Status' ]) validation_features = validation_data . drop ( columns = [ 'Status' ]) scaler . fit ( training_features ) # Perform the transformation normalized_training_data = scaler . transform ( training_features ) normalized_validation_data = scaler . transform ( validation_features ) # Replace old values with normalized values, keeping the 'Status' column intact training_data_normalized = pd . DataFrame ( normalized_training_data , columns = training_features . columns ) training_data_normalized [ 'Status' ] = training_data [ 'Status' ] validation_data_normalized = pd . DataFrame ( normalized_validation_data , columns = validation_features . columns ) validation_data_normalized [ 'Status' ] = validation_data [ 'Status' ] ( training_exploration , validation_exploration , training_data_normalized . head (), validation_data_normalized . head ()) ({'summary': Feature_1 Feature_2 Feature_3 Feature_4 Feature_5 Feature_6 \ count 300.000000 300.000000 300.000000 300.000000 300.000000 300.000000 mean 0.133587 0.353081 1.148927 3.317685 5.996621 3.559958 std 0.464645 0.683597 1.762382 4.925212 8.333159 5.400490 min -1.299247 -2.567137 -3.348178 -3.592233 -3.219585 -3.600452 25% -0.119510 -0.027414 0.086244 0.204398 0.364976 0.438351 50% 0.101041 0.249088 0.400961 0.584282 0.687151 0.770921 75% 0.425922 0.838214 2.208981 7.051022 17.101556 5.775406 max 1.437154 1.946377 6.346909 16.082052 20.141440 19.992030 Feature_7 Feature_8 Feature_9 Feature_10 ... Feature_492 \ count 300.000000 300.000000 300.000000 300.000000 ... 300.000000 mean 0.046032 -1.432604 -2.153879 -2.600541 ... -0.401728 std 2.784697 3.404206 3.952983 4.588838 ... 7.185068 min -4.989652 -7.164711 -8.709318 -9.939976 ... -18.490077 25% -1.386086 -5.368501 -7.316010 -8.702373 ... -2.815486 50% 0.473083 0.277924 -0.003062 0.095490 ... -0.825105 75% 0.856203 0.801420 0.864308 0.925883 ... 2.345776 max 19.567093 17.508389 4.250973 4.703203 ... 30.755470 Feature_493 Feature_494 Feature_495 Feature_496 Feature_497 \ count 300.000000 300.000000 300.000000 300.000000 300.000000 mean -0.394811 -0.445478 -0.450620 -0.396666 -0.354362 std 7.224344 7.078942 7.060402 7.041630 7.019486 min -18.238748 -18.861712 -18.839982 -19.969468 -18.730305 25% -2.990391 -3.295071 -3.132898 -2.849080 -3.417853 50% -0.752216 -0.628879 -0.571799 -0.546996 -0.424046 75% 2.197214 2.215800 1.815272 1.702486 1.530649 max 31.672199 32.124138 32.518974 33.409954 33.517280 Feature_498 Feature_499 Feature_500 Status count 300.000000 300.000000 300.000000 300.000000 mean -0.465986 -0.627828 -0.551134 2.000000 std 7.072625 7.298489 7.411132 0.817861 min -18.467430 -19.040604 -19.114384 1.000000 25% -3.537704 -3.896523 -4.135700 1.000000 50% -0.304221 -0.186471 -0.036254 2.000000 75% 1.866116 1.976870 1.988903 3.000000 max 33.039946 33.314707 32.928803 3.000000 [8 rows x 501 columns], 'missing_values': Feature_1 0 Feature_2 0 Feature_3 0 Feature_4 0 Feature_5 0 .. Feature_497 0 Feature_498 0 Feature_499 0 Feature_500 0 Status 0 Length: 501, dtype: int64, 'data_types': Feature_1 float64 Feature_2 float64 Feature_3 float64 Feature_4 float64 Feature_5 float64 ... Feature_497 float64 Feature_498 float64 Feature_499 float64 Feature_500 float64 Status float64 Length: 501, dtype: object}, {'summary': Feature_1 Feature_2 Feature_3 Feature_4 Feature_5 Feature_6 \ count 300.000000 300.000000 300.000000 300.000000 300.000000 300.000000 mean 0.173138 0.423545 1.257116 3.547109 6.052185 3.180508 std 0.405372 0.700782 1.850152 5.261167 8.478668 5.242294 min -1.463762 -2.487939 -1.737156 -2.285522 -3.048860 -2.582134 25% -0.036859 0.071475 0.175340 0.257864 0.415697 0.449917 50% 0.105102 0.269172 0.346894 0.479223 0.580994 0.666464 75% 0.463262 0.861609 2.473455 7.796831 17.104935 4.728514 max 1.128735 2.118145 6.726057 17.136014 19.962632 19.762905 Feature_7 Feature_8 Feature_9 Feature_10 ... Feature_492 \ count 300.000000 300.000000 300.000000 300.000000 ... 300.000000 mean -0.365900 -1.780681 -2.394442 -2.817422 ... -0.790263 std 2.354452 3.154211 3.949272 4.512745 ... 7.807457 min -5.187750 -7.439548 -8.746702 -9.616524 ... -28.605543 25% -1.874546 -5.506959 -7.375137 -8.745140 ... -3.018682 50% 0.493582 -0.078736 -0.261905 -0.316128 ... -0.826995 75% 0.746429 0.768297 0.839751 0.887290 ... 3.014377 max 13.640237 3.371212 3.305640 3.391341 ... 19.788439 Feature_493 Feature_494 Feature_495 Feature_496 Feature_497 \ count 300.000000 300.000000 300.000000 300.000000 300.000000 mean -0.659900 -0.490028 -0.449033 -0.461000 -0.373395 std 7.995240 8.092395 8.208499 8.284201 8.071444 min -28.988463 -29.593378 -30.162245 -30.913387 -31.111843 25% -3.597136 -3.871761 -3.723010 -3.046614 -3.784670 50% -0.746635 -0.685207 -0.575542 -0.466373 -0.374921 75% 3.570288 3.665994 4.371823 4.082067 4.203521 max 19.820185 20.020610 21.382359 21.651101 21.274513 Feature_498 Feature_499 Feature_500 Status count 300.000000 300.000000 300.000000 300.000000 mean -0.336339 -0.580490 -0.790558 2.000000 std 8.010108 8.018548 8.015539 0.817861 min -31.307439 -31.974854 -32.149645 1.000000 25% -3.448527 -3.160960 -3.611293 1.000000 50% -0.257718 -0.143136 -0.019509 2.000000 75% 3.663612 2.905283 2.545813 3.000000 max 21.593234 21.266144 21.462606 3.000000 [8 rows x 501 columns], 'missing_values': Feature_1 0 Feature_2 0 Feature_3 0 Feature_4 0 Feature_5 0 .. Feature_497 0 Feature_498 0 Feature_499 0 Feature_500 0 Status 0 Length: 501, dtype: int64, 'data_types': Feature_1 float64 Feature_2 float64 Feature_3 float64 Feature_4 float64 Feature_5 float64 ... Feature_497 float64 Feature_498 float64 Feature_499 float64 Feature_500 float64 Status float64 Length: 501, dtype: object}, Feature_1 Feature_2 Feature_3 Feature_4 Feature_5 Feature_6 \ 0 0.046721 0.450267 0.583346 0.598182 1.214287 2.325745 1 1.028993 0.936947 0.866022 0.875283 1.469665 1.812035 2 2.262543 0.478249 1.034278 1.320238 1.673551 1.017062 3 1.036933 0.431232 0.469079 0.384968 1.108933 2.491662 4 0.582966 0.451415 0.797502 0.798521 1.471966 1.814317 Feature_7 Feature_8 Feature_9 Feature_10 ... Feature_492 Feature_493 \ 0 0.512322 -0.971569 -1.282061 -1.304200 ... -0.014684 -0.274264 1 -0.304147 -1.224078 -1.304439 -1.344016 ... 0.765367 -0.119435 2 -0.915923 -1.366142 -1.332514 -1.376258 ... -1.189046 -1.821036 3 0.886277 -0.917163 -1.176759 -1.222860 ... -1.697818 -1.386824 4 -0.212359 -1.156642 -1.356878 -1.361469 ... -0.054246 -0.180881 Feature_494 Feature_495 Feature_496 Feature_497 Feature_498 \ 0 -0.649606 -1.672951 -2.356248 -0.931145 0.196579 1 -0.535273 -0.864003 -1.341570 -1.734005 -1.285506 2 -1.553779 -0.726802 -0.236522 -0.204568 -0.331442 3 -0.735262 -0.500159 -0.533523 -0.780537 -1.417013 4 -0.419657 -0.907792 -1.786335 -1.955934 -0.890490 Feature_499 Feature_500 Status 0 0.551889 0.896501 1.0 1 -0.651659 -0.376686 1.0 2 -0.690672 -1.511005 1.0 3 -1.716246 -1.244506 1.0 4 -0.163793 0.018296 1.0 [5 rows x 501 columns], Feature_1 Feature_2 Feature_3 Feature_4 Feature_5 Feature_6 \ 0 -0.022650 0.743993 1.150584 1.342328 1.665202 1.027499 1 1.330917 1.855621 2.273000 2.162092 1.451855 0.035117 2 1.216750 1.642358 1.820092 1.898458 1.579003 0.361406 3 0.353781 0.287302 0.389961 0.540607 1.254611 2.347723 4 1.305979 0.600910 0.850079 0.983304 1.574349 1.526335 Feature_7 Feature_8 Feature_9 Feature_10 ... Feature_492 Feature_493 \ 0 -0.849064 -1.382491 -1.438760 -1.422004 ... -1.621809 -2.186896 1 -1.445017 -1.533140 -1.583451 -1.507165 ... 0.428643 0.790528 2 -1.324071 -1.568474 -1.524527 -1.468575 ... -0.463006 -0.442640 3 0.532883 -1.030196 -1.295744 -1.278027 ... 0.459729 1.064713 4 -0.420408 -1.236318 -1.324502 -1.369341 ... -0.405411 -0.612238 Feature_494 Feature_495 Feature_496 Feature_497 Feature_498 \ 0 -0.998012 -0.022176 0.323118 0.586686 1.045629 1 1.486304 2.068227 1.496972 0.647572 0.292991 2 -0.670761 -1.201971 -1.802570 -1.549445 -0.730685 3 2.182513 2.071373 0.569010 -0.154382 -0.440969 4 -1.243341 -1.826906 -1.541966 -0.722757 -0.304865 Feature_499 Feature_500 Status 0 1.702677 1.903927 1.0 1 0.309244 0.416158 1.0 2 -0.304573 -0.293324 1.0 3 -0.741868 -1.319178 1.0 4 -0.237051 -0.394179 1.0 [5 rows x 501 columns]) Model Development Develop two machine learning models: Model A: Using the original feature set. Model B: Using the enhanced feature set from advanced feature extraction. Explore various modeling techniques for both models. from sklearn.ensemble import RandomForestClassifier from sklearn.svm import SVC from sklearn.neural_network import MLPClassifier from sklearn.metrics import accuracy_score , precision_score , recall_score , f1_score from sklearn.model_selection import train_test_split # Split the training data into features and target variable X_train = training_data_normalized . drop ( 'Status' , axis = 1 ) y_train = training_data_normalized [ 'Status' ] # Split the validation data into features and target variable X_valid = validation_data_normalized . drop ( 'Status' , axis = 1 ) y_valid = validation_data_normalized [ 'Status' ] # Initialize the models rf_clf = RandomForestClassifier ( random_state = 0 ) svm_clf = SVC ( random_state = 0 ) mlp_clf = MLPClassifier ( random_state = 0 ) # Train the models on the training data rf_clf . fit ( X_train , y_train ) svm_clf . fit ( X_train , y_train ) mlp_clf . fit ( X_train , y_train ) # Predict on the validation data rf_preds = rf_clf . predict ( X_valid ) svm_preds = svm_clf . predict ( X_valid ) mlp_preds = mlp_clf . predict ( X_valid ) # Evaluate the models models = [ 'Random Forest' , 'SVM' , 'Neural Network' ] predictions = [ rf_preds , svm_preds , mlp_preds ] # Function to calculate metrics def evaluate_model ( y_true , y_pred ): accuracy = accuracy_score ( y_true , y_pred ) precision = precision_score ( y_true , y_pred , average = 'macro' ) recall = recall_score ( y_true , y_pred , average = 'macro' ) f1 = f1_score ( y_true , y_pred , average = 'macro' ) return accuracy , precision , recall , f1 In [1]: In [2]: In [3]: In [4]: In [5]: In [6]: In [7]: In [8]: Out[8]: In [9]: In [10]: In [11]: In [12]: Out[12]: In [13]: