Skip to content

Latest commit

 

History

History
82 lines (60 loc) · 4.29 KB

README.md

File metadata and controls

82 lines (60 loc) · 4.29 KB

Loan-Prediction-SVM

Predicting Loan State with SVM method, preprocessing and feature selction

Tech 🛠️ Languages and Tools :

Python  Jupyter Notebook  Google Colab  Numpy  Pandas  MatPlotLib  seaborn  Sci-kit Learn 

Run the Notebook on Google Colab

You can easily run this code on google colab by just clicking this badge Open In Colab

Algorithms used

  • Pearson Correlation : for features selection
  • SVM algorithm : for training model
  • Grid Search : for finding best hyper parameters

About the Dataset

Download Dataset

you can use this Dataset with clicking this badges :

Train set : Static Badge

Test Set : Static Badge

Using Dataset on jupyter notebook

on the step 3 of the code file you can easily use the dataset like importing this code:

train_set_url = "https://raw.githubusercontent.com/AsadiAhmad/Loan-Prediction-SVM/refs/heads/main/Dataset/train.csv"
test_set_url = "https://raw.githubusercontent.com/AsadiAhmad/Loan-Prediction-SVM/refs/heads/main/Dataset/test.csv"

pd.set_option('display.max_rows', None)

train_set = pd.read_csv(train_set_url)
test_set = pd.read_csv(test_set_url)

Dataset Structure

Column Description Type
Id Unique Loan ID Int
Income The income person have Int
Age Age of person Int
Experience No. of years of experience Int
Married/Single Married/Single state : single/married String
House_Ownership House ownership : owned/rented/norent_noown String
Car_Ownership Car ownership : yes/no String
Profession The profession person have String
CITY The city person live String
STATE The state person live String
CURRENT_JOB_YRS How many years the person have the job Int
CURRENT_HOUSE_YRS How many years the person have the house Int
Risk_Flag Loan State: 0/1 Int

Execution time

For running all sell on hosted runtime it costs about 65 minutes (around 1 hours) but there is tips to run it around 15 minutes.

But you can just using one kernel for the SVM so here are more details:

  • Linear Kernel : 2 min
  • Poly Kernel : 2 min
  • RBF Kernel : 3 min
  • Sigmoid Kernel : 2 min

and for the finding the optimized hyperparameter you need to run again this kernels but you can optimize the hyperparameters (Step 10) and skip first training model (Step 9).

the kernels are not have different accuracy the best one is rbf model so run that in the** Step 10** (finding hyperparameter) running time in google colab is just around 15 min.

License

This project is licensed under the MIT License.