Methodologies
6 min
mantacore methodology overview introduction mantacore is a comprehensive financial analysis and portfolio management system designed to provide robust tools for data cleaning, risk modeling, portfolio optimization, and performance attribution this document provides a high level overview of the methodologies implemented in the system, serving as an introduction to the more detailed documentation available for each component for detailed information on specific methodologies, please refer to the dedicated documentation for each component key methodologies data cleaning the data cleaning process ensures high quality, consistent data for financial analysis by removing outliers and handling missing values aligning time series data across different instruments rescaling bond prices according to time to maturity providing different strategies for handling data anomalies for more details, see data cleaning process for linear instruments docid\ buc0mi9ohsl898fnc45zq risk factor modeling the risk factor engine analyzes the underlying factors that drive financial instrument returns linear risk factor models for standard instruments regression based analysis to identify factor exposures risk decomposition to understand sources of risk factor based scenario analysis for more details, see risk factor modeling methodology docid\ jptyxhvypbehcbsbcervw portfolio construction / optimization the optimization engine constructs optimal portfolios based on mean cvar optimization risk based allocation methods constraint based optimization with various objective functions efficient frontier generation for risk return analysis for more details, see portfolio construction docid\ yjp6dberjogs yzae1kit performance attribution the attribution system analyzes portfolio performance brinson fachler attribution based on a holdings based attribution type hierarchical performance attribution based on a returns based attribution type risk & performance decomposition across different dimensions for more details, please see performance attribution docid\ gs 4wlffhfsk0ficbo2s7