Please enable JavaScript.
Coggle requires JavaScript to display documents.
Financial Fraud Detection using Outlier Detection Technique, Natural…
Financial Fraud Detection using Outlier Detection Technique
Word Embedding Techniques
(Enriching Feature Vectors)
Word2Vec
Doc2Vec
Bert(Bidirectional Encoder Representations From Transformers)
Challenges
Limitation of Financial statement dataset
Insufficient Data Size
Fraudulent Companies Financial Statements & Financial Ratios
Non-Fraudulent Companies Financial Statements & Financial Ratios
Imbalanced Dataset
Few Fraudulent Financial Statements
Data
Structured
Non-Financial Ratios
Financial statements(balance sheet, income statement and cash-flow)
Financial Ratios
Financial statements(balance sheet, income statement and cash-flow)
Liquidity
Leverage
Efficiency
Profitability
Market Value
Unstructured
Textual Content
Financial Statements (Management Discussion
and Analysis Section
Financial social media platform
Conference call transcripts
Vocal Speech
Earnings Conference Calls
Reference Databases
Detecting companies fraud status
Stock exchange
SEC's AAER
SEC & Fut Center
TEJ
Audit Analytics
Compustat
Seeking Alpha
CSRC
Lebanese gov.institution
Unsupervised Classifiers
Outlier Detectors
Unsupervised Model
Affinity Propagation
DBScan
Self-Organizing Map
Isolation Forest
Data Sources
Financial Statement Ratio/Variable
MD & A section
Financial Social Media Platform
Conference Call Transcript
Earnings Conference Call
Data Source by Country
US
Greece
Taiwan
South Africa
Indonesia
India
Lebanon
Vietnam
Korea
Taipei
Future Work
Usage of bio-inspired algorithms
artificial immune systems (AIS),
genetic algorithms (GA)
Enriching feature vectors
Data size
Imbalanced dataset
Model Building Process
Include Social media data analysis
Unstructured Data
Vocal Speech
Natural Language Processing