the prediction of survival of hepatitis patient

1
HepatitisThe prediction of survival for the
hepatitis patientGroup No7
No Regno name course
1 T/UDOM/2020/
2 T/UDOM/2020/
3 T/UDOM/2020/
4 T/UDOM/2020/
2
Insight on hepatitis

• Hepatitis means inflammation of the liver. The
  liver is a vital organ that processes nutrients,
  filters the blood, and fights infections.
• according to the centers for disease control and
  prevention(CDC)Trusted Sources there are
  approximately 4.4 million Americans are currently
  living with chronic hepatisis b and c.many people
  dont even know that they have hepatitis

3
Problem statement

• It is hard to predict the chances of survial of a
  patient with hepatitis
• objective
• To use machine learning algorithms to find or to
  determine the chance of survival of the patient
  with hepatitis

4
Methodology

• By using machine learning algorithms with python
  programming language
• By using hepatitis dataset from uci data set
  repository

5
Importing libraries

• Import pandas,numpy,seaborn and matplotlib.pyplot

6
Loading data from csv file
7
Data description for the complete instances
8
Data set cleaning

• As we see our data set is not clear there are
  some values are missed denoted by ?
• Also our data are objective data type and some
  attributes have imbalanced data
• Data is not normalized
• So what we do?
• Replacing the missing values by the mean value
  for the numerical value and replacing by model
  value for the categorical values
• Changing the whole data frame to float data type
  and using smote,standard scaller and normalization

9
Mean calculation
replacements
10
Data description after replacing the missing
values and after changing data to float data type
11
Data visualization for some attributes to show
skewness
Our data are not normalized
12
Normalization By using np.log

• hep_replace'ALBUMIN', 'ALK PHOSPHATE',
  'BILIRUBIN', 'SGOT' hep_replace'ALBUMIN','AL
  K PHOSPHATE', 'BILIRUBIN', 'SGOT'.applymap(np.lo
  g)

13
Feauture selection

• We selected all attributes for training testing
  and prediction but our target is class attribute

14
Checking for imbalanced data
15
By using smote to balance our data set
16
Standard scaler
17
Training testing and predictions

• By using LogisticRegression ,GaussianNB,DecisionTr
  eeClassifier for training testing and prediction
  of our dataset

18
Rogisticregression predictions ad score
19
Logisticregression classification report and its
confusion matrix
20
logisticregression

• Comfusion matrix

21
Tuning algorithm performance by ensembles

• Using Random forest classfier

22
Predictions by Randomforestclassfier and its
confusion matrix
23
Confusion matrix for the .
24
conclusion

• With randomforestclassfier we got an accuracy of
  92

25
More Thanks