This chapter will first give a formal definition of fraud. You will then learn how to detect anomalies in the type of payment methods used or the time these payments are made to flag suspicious transactions.

Introduction & Motivation

Imbalanced class distribution

Cost of not detecting fraud

Time features

Circular histogram

Suspicious timestamps

Frequency features

Frequency feature for one account

Frequency feature for multiple accounts

Recency features

Recency feature

Comparing frequency & recency

In the second chapter, you will learn how to use networks to fight fraud. You will visualize networks and use a sociology concept called homophily to detect fraudulent transactions and catch fraudsters.

Social network analytics

Analyzing a network

Overlapping edges

Fraud and social network analysis

Looking for homophily in a network

Visualizing node attributes

Social network based inference

Relational vs non-relational models

Relational neighbor classifier

Social network metrics

Degree, closeness & betweenness

Adding network features

Fortunately, fraud occurrences are rare. However, this means that you're working with imbalanced data, which if left as is will bias your detection models. In this chapter, you will tackle imbalance using over and under-sampling methods.

Dealing with imbalanced datasets

How to deal with class imbalance?

Visualizing patterns in the data

Random over-sampling

Random under-sampling

Shrinking the majority group

Combining ROS & RUS

Synthetic Over-sampling

Have you met SMOTE?

SMOTE

From dataset to detection model

Build your own detection model

True cost of fraud detection

Imbalanced class distributions

In this final chapter, you will learn about a surprising mathematical law used to detect suspicious occurrences. You will then use robust statistics to make your models even more bulletproof.

Analisis digit menggunakan hukum Benford

Hukum Benford untuk digit pertama

Kesesuaian data sensus

Hukum Benford untuk deteksi kecurangan

Kesesuaian dengan Hukum Benford

Klaim asuransi kebakaran

Himpunan data pembayaran

Mendeteksi pencilan univariat

Menghitung robust z-score

Boxplot

Mendeteksi pencilan multivariat

Deteksi outlier multivariat

Digit analysis and robust statistics

Chapter 1 datasets

Chapter 2 datasets

Chapter 3 datasets

Chapter 4 datasets

Association of Certified Fraud Examiners memperkirakan bahwa kecurangan menelan biaya bagi organisasi di seluruh dunia sebesar $3,7 triliun per tahun dan perusahaan rata-rata kehilangan lima persen dari pendapatan tahunannya akibat fraud. Upaya kecurangan bahkan diperkirakan akan terus meningkat di masa depan, sehingga deteksi fraud menjadi sangat diperlukan di sebagian besar industri. Kursus ini akan menunjukkan bagaimana mempelajari pola fraud dari data historis dapat digunakan untuk memerangi kecurangan. Beberapa teknik dari statistika robust dan analisis digit disajikan untuk mendeteksi pengamatan tidak biasa yang kemungkinan terkait dengan fraud. Dua tantangan utama saat membangun alat terawasi untuk deteksi fraud adalah ketidakseimbangan atau kemencengan data dan beragamnya biaya untuk berbagai jenis salah klasifikasi. Kami menyajikan teknik untuk mengatasi isu-isu ini dan berfokus pada himpunan data buatan dan nyata dari beragam aplikasi fraud.

Unsupervised Learning in R

Supervised Learning in R: Classification

Pelajari deteksi penipuan dengan deteksi anomali, visualisasi jaringan sosial, dan analisis digit di R.

Deteksi Fraud di R

Pelajari cara mendeteksi penipuan dengan analisis data menggunakan R.

Himpunan data pembayaran

Deteksi Fraud di R

Latihan interaktif praktis