Please enable JavaScript.
Coggle requires JavaScript to display documents.
Reading 57: Fintech in Investment Management - Coggle Diagram
Reading 57: Fintech in Investment Management
Definition
Fintech: Financial technology, refers to technological developments that are or may be applied to the financial service industry
Fintech companies are firms that develop these technologies
Big Data
Traditional sources: securities markets, financial statements, regulatory filings, economic statistics
Nontraditional sources:
Individuals (social media, online reviews, website vitsits, etc.)
Corporate exhaust (e.g., retail scanner data)
Internet of things (e.g., smart building)
Volume: Gigabyte → Terabyte → Petabyte
Velocity: ranges from low latency (real-time data) to high latency (periodic reports)
Variety: ranges from structured (e.g., spreadsheet, database) to semi-structured (e.g., HTML code) to unstructured (video)
AI
Artificial intelligence: refers to computer systems that can learn and make decisions in a manner similar to the human brain
Machine learning algorithms
Learn to make decisions based on historical data
Identify structure and patterns without human help
Require Big Data to train and validate ML models
Types of learning
Unsupervised learning: Detect and recognized patterns in input dataset
Deep learning: Use layers of neural networks to detect increasingly complex patterns, may use supervised or unsupervised learning
Supervised learning: Model output dataset based on input data set
Problems
Overfitting: Machine produces too complex of a model, treats noise as true parameters
Underfitting: Machine produces too simple of a model, treats true parameters as noise
Investment Management Application
Text analytics: Analyze unstructured voice or text (e.g., interpret regulatory filings)
Natural language processing: (e.g., discern sentiment from research reports)
Risk governance (e.g., apply machine learning to scenario analysis in stress testing)
Algorithmic trading (e.g., high-frequency trading to profit from mispricings based on tick data)
Robo-advisory services
Automated investment advice based on client's answers to questions about financial position, risk tolerance, etc.
Pro: low cost to clients
Cons: Clients may not understand the reasoning behind recommendations
Distributed ledger:
Offer almost real-time trade verification, reconciliation, and settlement
Thus streamline the labor-intensive post-trade process
Blockchain: Distributed ledger that records transactions sequentially
Net work can be permission-less (all users have equal access) or permissioned (different level of access)
Tokenization
represent ownership rights of physical assets on a blockchain or distributed ledger → to verify ownership, authenticity, historical activities