Create a Linear Classifier Model in 5 Steps using Tensorflow Real World Data Set

Step 1:¶

import modules¶

In [4]:

import tensorflow as tf

import numpy as np
import pandas as pd

import matplotlib.pyplot as plt

from sklearn.model_selection import train_test_split
from sklearn.metrics import confusion_matrix,classification_report


%matplotlib inline

Step 2:¶

process data¶

In [5]:

def Data_Process():
    
    """
    This will read the CSV and Normalize the Data and
    Perform Train Test Split and Return
    X_Train, X_Test, Y_Train, Y_Test
    
    """
    # Name for the column  or Features Map
    columns_to_named = ["Pregnancies","Glucose","BloodPressure",
           "SkinThickness","Insulin","BMI","DiabetesPedigreeFunction",
           "Age","Class"]
    
    # Read the Dataset and Rename the Column
    df = pd.read_csv("pima-indians-diabetes.csv",header=0,names=columns_to_named)

    col_norm =['Pregnancies', 'Glucose', 'BloodPressure', 'SkinThickness', 'Insulin',
       'BMI', 'DiabetesPedigreeFunction']
    
    # Normalization using Custom Lambda Function
    
    df1_norm = df[col_norm].apply(lambda x :( (x - x.min()) / (x.max()-x.min()) ) )
    
    X_Data = df1_norm
    Y_Data = df["Class"]
    
    X_Train, X_Test, Y_Train, Y_Test = train_test_split(X_Data,Y_Data, test_size=0.3,random_state=101)
    
    return X_Train, X_Test, Y_Train, Y_Test

Step 3:¶

Define Feature Columns¶

In [6]:

def create_feature_column():
    
    feat_Pregnancies = tf.feature_column.numeric_column('Pregnancies')
    feat_Glucose = tf.feature_column.numeric_column('Glucose')
    feat_BloodPressure = tf.feature_column.numeric_column('BloodPressure')
    feat_SkinThickness_tricep = tf.feature_column.numeric_column('SkinThickness')
    feat_Insulin = tf.feature_column.numeric_column('Insulin')
    feat_BMI = tf.feature_column.numeric_column('BMI')
    feat_DiabetesPedigreeFunction  = tf.feature_column.numeric_column('DiabetesPedigreeFunction')
    
    feature_column = [feat_Pregnancies, feat_Glucose, feat_BloodPressure, 
                  feat_SkinThickness_tricep, feat_Insulin, 
                 feat_BMI , feat_DiabetesPedigreeFunction] 
    
    return feature_column

Create Input Function and Test Function¶

In [7]:

X_Train, X_Test, Y_Train, Y_Test = Data_Process()
feature_column = create_feature_column()

input_func = tf.estimator.inputs.pandas_input_fn(x=X_Train, y=Y_Train,
                                                 batch_size=40,num_epochs =1000, 
                                                 shuffle=True)

eval_input_func = tf.estimator.inputs.pandas_input_fn(x=X_Test,
                                                      y=Y_Test,
                                                      batch_size=40,
                                                      num_epochs=1,
                                                      shuffle=False)

Step 4 :¶

Create Linear Classifier Model¶

In [8]:

model = tf.estimator.LinearClassifier(feature_columns=feature_column, 
                                      n_classes=2)

INFO:tensorflow:Using default config.
WARNING:tensorflow:Using temporary folder as model directory: /var/folders/yh/7gktt0ls0fj77fnrs694ht6m0000gn/T/tmp6ydp5d48
INFO:tensorflow:Using config: {'_model_dir': '/var/folders/yh/7gktt0ls0fj77fnrs694ht6m0000gn/T/tmp6ydp5d48', '_tf_random_seed': None, '_save_summary_steps': 100, '_save_checkpoints_steps': None, '_save_checkpoints_secs': 600, '_session_config': allow_soft_placement: true
graph_options {
  rewrite_options {
    meta_optimizer_iterations: ONE
  }
}
, '_keep_checkpoint_max': 5, '_keep_checkpoint_every_n_hours': 10000, '_log_step_count_steps': 100, '_train_distribute': None, '_device_fn': None, '_protocol': None, '_eval_distribute': None, '_experimental_distribute': None, '_service': None, '_cluster_spec': <tensorflow.python.training.server_lib.ClusterSpec object at 0x1a308afe48>, '_task_type': 'worker', '_task_id': 0, '_global_id_in_cluster': 0, '_master': '', '_evaluation_master': '', '_is_chief': True, '_num_ps_replicas': 0, '_num_worker_replicas': 1}

Step 5:¶

Train¶

In [10]:

history = model.train(input_fn=input_func, steps = 1000)

INFO:tensorflow:Calling model_fn.
INFO:tensorflow:Done calling model_fn.
INFO:tensorflow:Create CheckpointSaverHook.
INFO:tensorflow:Graph was finalized.
WARNING:tensorflow:From /anaconda3/lib/python3.7/site-packages/tensorflow/python/training/saver.py:1266: checkpoint_exists (from tensorflow.python.training.checkpoint_management) is deprecated and will be removed in a future version.
Instructions for updating:
Use standard file APIs to check for files with this prefix.
INFO:tensorflow:Restoring parameters from /var/folders/yh/7gktt0ls0fj77fnrs694ht6m0000gn/T/tmp6ydp5d48/model.ckpt-5000
WARNING:tensorflow:From /anaconda3/lib/python3.7/site-packages/tensorflow/python/training/saver.py:1070: get_checkpoint_mtimes (from tensorflow.python.training.checkpoint_management) is deprecated and will be removed in a future version.
Instructions for updating:
Use standard file utilities to get mtimes.
INFO:tensorflow:Running local_init_op.
INFO:tensorflow:Done running local_init_op.
INFO:tensorflow:Saving checkpoints for 5000 into /var/folders/yh/7gktt0ls0fj77fnrs694ht6m0000gn/T/tmp6ydp5d48/model.ckpt.
INFO:tensorflow:loss = 22.763405, step = 5001
INFO:tensorflow:global_step/sec: 341.279
INFO:tensorflow:loss = 18.86869, step = 5101 (0.294 sec)
INFO:tensorflow:global_step/sec: 277.223
INFO:tensorflow:loss = 20.387632, step = 5201 (0.363 sec)
INFO:tensorflow:global_step/sec: 372.544
INFO:tensorflow:loss = 18.063845, step = 5301 (0.269 sec)
INFO:tensorflow:global_step/sec: 472.739
INFO:tensorflow:loss = 15.5112705, step = 5401 (0.210 sec)
INFO:tensorflow:global_step/sec: 562.262
INFO:tensorflow:loss = 18.075052, step = 5501 (0.180 sec)
INFO:tensorflow:global_step/sec: 562.68
INFO:tensorflow:loss = 21.951363, step = 5601 (0.178 sec)
INFO:tensorflow:global_step/sec: 318.397
INFO:tensorflow:loss = 20.899546, step = 5701 (0.312 sec)
INFO:tensorflow:global_step/sec: 331.633
INFO:tensorflow:loss = 19.533257, step = 5801 (0.301 sec)
INFO:tensorflow:global_step/sec: 428.697
INFO:tensorflow:loss = 21.393614, step = 5901 (0.235 sec)
INFO:tensorflow:Saving checkpoints for 6000 into /var/folders/yh/7gktt0ls0fj77fnrs694ht6m0000gn/T/tmp6ydp5d48/model.ckpt.
INFO:tensorflow:Loss for final step: 19.936054.

Test¶

In [16]:

results = model.evaluate(eval_input_func)

INFO:tensorflow:Calling model_fn.
WARNING:tensorflow:Trapezoidal rule is known to produce incorrect PR-AUCs; please switch to "careful_interpolation" instead.
WARNING:tensorflow:Trapezoidal rule is known to produce incorrect PR-AUCs; please switch to "careful_interpolation" instead.
INFO:tensorflow:Done calling model_fn.
INFO:tensorflow:Starting evaluation at 2019-05-17T16:07:31Z
INFO:tensorflow:Graph was finalized.
INFO:tensorflow:Restoring parameters from /var/folders/yh/7gktt0ls0fj77fnrs694ht6m0000gn/T/tmp6ydp5d48/model.ckpt-6000
INFO:tensorflow:Running local_init_op.
INFO:tensorflow:Done running local_init_op.
INFO:tensorflow:Finished evaluation at 2019-05-17-16:07:31
INFO:tensorflow:Saving dict for global step 6000: accuracy = 0.74458873, accuracy_baseline = 0.64935064, auc = 0.7916461, auc_precision_recall = 0.6702014, average_loss = 0.52323383, global_step = 6000, label/mean = 0.35064936, loss = 20.144503, precision = 0.7037037, prediction/mean = 0.3514547, recall = 0.4691358
INFO:tensorflow:Saving 'checkpoint_path' summary for global step 6000: /var/folders/yh/7gktt0ls0fj77fnrs694ht6m0000gn/T/tmp6ydp5d48/model.ckpt-6000

In [20]:

results["accuracy"]

Out[20]:

0.74458873