AI 용어집
인공지능 완전 사전
Alpha Alternatif
Excess performance generated by unconventional strategies using alternative data, distinct from traditional alpha derived from conventional financial markets.
Backtesting sur Données Alternatives
Process of validating a quantitative strategy by simulating it on historical alternative data to evaluate its performance and robustness before deployment.
Crowdsourcing Alpha
Method of collecting and aggregating investment signals or analyses from a wide community of experts to generate alpha in quantitative strategies.
Données de Géolocalisation
Information derived from the spatial location of mobile devices, used to anticipate foot traffic at retail points or consumption trends.
Données de Crédit Carte
Anonymized and aggregated credit card transactions, providing real-time indicators on household spending and the health of economic sectors.
Factor Investing Alternatif
Factor investing approach based on performance factors derived from alternative data, such as social media sentiment or construction activity.
Natural Language Processing (NLP) Financier
Set of AI techniques applied to text analysis (reports, news, social media) to extract quantitative signals and sentiment indicators.
Nowcasting
Prediction of economic or financial variables in near real-time using alternative data streams, bypassing the delay of official statistics.
Satellite Imagery Analysis
Satellite imagery analysis to estimate economic activity, such as counting cars in supermarket parking lots or tracking container ships.
Sentiment Analysis
Natural language processing technique aimed at quantifying the polarity (positive, negative, neutral) of a text to derive a market signal.
Alternative Signal-to-Noise Ratio (SNR)
Crucial metric evaluating the quality of an investment signal from alternative data, comparing its predictive power to its intrinsic volatility.
Financial Web Scraping
Automated extraction of unstructured data from websites (prices, forums, job postings) to transform them into actionable quantitative indicators.
Reinforcement Learning for Trading
Use of reinforcement learning algorithms where an agent learns an optimal trading policy through trial and error to maximize a reward function.
ESG Data Integration
Incorporation of Environmental, Social, and Governance data, often unstructured and alternative, into quantitative asset valuation models.
Overfitting on Alternative Data
Major risk where a quantitative model learns the specific noise of an alternative data set rather than the underlying signal, leading to poor performance in production.
Data Vending
Practice of monetizing raw or enriched alternative data sets, specifically created and sold to quantitative funds.