From eca2c43bbfedf74da92362d730393f3ca84a752c Mon Sep 17 00:00:00 2001 From: ShayanHodai Date: Fri, 31 May 2024 23:10:58 -0400 Subject: [PATCH] Update README.md --- README.md | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/README.md b/README.md index d2dc498..9bc570e 100644 --- a/README.md +++ b/README.md @@ -12,7 +12,7 @@ The dataset is highly imbalanced as, less than 1% of total transactions are frau ![Example Image](images/imbalanced%20dataset.png) features histograms: as seen, most features are centred around 0 -[View the PDF](images/features\histogram.pdf) +[View the PDF](images/features%20histogram.pdf) Time and Amount features need scaling Time feature is scaled by StandardScaler -> range between 0 to 1 @@ -25,13 +25,13 @@ correlation of fraud/normal transactions with non-redundant features Machine learning classification models evaluation metrics: As the cost of False Positive and False Negative in this problem varies, Precision and Recall and, eventually, f1-score are the evaluation metrics of the model performance Logistic regression: -![Example Image](images/logistic\regression.png) +![Example Image](images/logistic%20regression.png) KNN: ![Example Image](images/KNN.png) SVM: ![Example Image](images/SVM.png) Decision tree classifier, which is highly overfitting: -![Example Image](images/Decision\Tree.png) +![Example Image](images/Decision%20Tree.png) ROC carve: ![Example Image](images/ROC.pdf) @@ -40,4 +40,4 @@ Fine-tuning the best performing model, which is logistic regression: ![Example Image](images/fine-tuning.png) Evaluation on the test set: -![Example Image](images/evaluation\on\test.png) +![Example Image](images/evaluation%20on%20test.png)