NETFLIX-APPTENCY

The test set. it consists of everything in train.csv except target.

Test set csv - https://www.kaggle.com/competitions/netflix-appetency/data?select=test.csv

The training set. it consists of an id column, the customers features, and a target column: target.

Train set csv - https://www.kaggle.com/competitions/netflix-appetency/data?select=train.csv

A sample submission file in the correct format target=1 means that the customer subscribes to Netflix

Sample_submission - https://www.kaggle.com/competitions/netflix-appetency/data?select=sample_submission.csv

steps:-

1. IMPORTS

• NUMPY
• PANDAS
• SEABORN
• MATPLOTLIB.PYPLOT
• DATETIME
• SKLEARN.PREPROCESSING
• CATBOOST
• XGBOOST
• LIGHTgbm
• sklearn.preprocesing(LabelEncoder)
• sklearn.model_selection(cross_val_score, Kfold, RepeatedStratifiedKFold, StratifiedKFold, cross_val_predict)
• sklearn.model_selection(train_test_split, GridSearchCV)
• sklearn.metrics(roc_auc_score)

2 IMPORT AND READ THE DATASET

• TRAIN.CSV - the training set. it consists of an id column, the customers features, and a target column: target.
• TEST.CSV - the test set. it consists of everything except target.

NOTE:- USED ----> (%config Completer.use_jedi = False ) MAGIC COMMAND. - Once you have enabled and run the %config Completer.use_jedi = False magic command, you can trigger the code autocompletion by pressing the tab button after the "." character.

3. ANALYSIS

• HEAD()

• SHAPE()

• DESCRIBE() ----> FIND MISSING DATA:-

• get_mising- function made to find missing data

• A histplot to show the distribution of missing data

• A histplot to show the distribution of missing data in Missing_df

  • To get columns with more than 25% missing values.

• Drop them from test and train set.

• Print categorical features of datatype object.

• Print numerical features of datatype object.

• Plot a pichart to show the percentage of numeric and categorical features.

• Fill with median(Numerical Features).

• Fill with mode(Categorical Features).

• Find columns that contain date objects.

• Apply datetime format.

• Show datetime features.

• Get each part of datetime using pandas DatetimeIndex.

• Drop from train/test.

• Update Categorical_Features list.

• Create a copy of datasets.

  • For train and test.

• Get the number of unique values for each feature.

• Categorical Features Sorted by cardinality.

STEPS to deal with missing data

• create a function "get_missing" to find the missing data
• store missing values from df_train in a dataframe "Missing_Df"
• Make a plot to visualize the missing values in "df_train".
• Create a dataframe "Missing_custom" to drop the missing columns with - - percentage greater than 25