Software for Stock Market Prediction
4 likes1,555 views
The document presents a software model for stock market prediction developed by Alexander Makarenko and Valeriy Soliya, detailing four types of predictive models based on broker reputation dynamics. It outlines the advantages of the model, including its predictive power and applicability to real-world broker interactions, and specifies various input parameters and data formats essential for effective modeling. Additionally, it provides insights into the software interfaces designed for different modeling approaches.
Software for Stock Market Prediction
- 1. <Эмблема> SOFTWARE FOR STOCK MARKET PREDICTION Alexander MAKARENKO, Prof., Valeriy SOLIYA, Institute of Applied System Analysis at National Technical University of Ukraine (KPI), 37, Pobedy Avenue, 03056, Kiev- 56, UKRAINE makalex@i.com.ua
- 2. MARKET PREDICTION (CONTENT) Model We represent Model Variants the learning Advantage software for Input parameters modeling and Input data formats prediction of Results of modeling stock market Interface behavior
- 3. MARKET PREDICTION (MODEL) ORIGINAL MODEL FOR LARGE SOCIAL, ECONOMICAL, POLITICAL, SYSTEMS PROPOSED BY prof. A.MAKARENKO E-mail: makalex@i.com.ua ; makalex@mmsa.ntu-kpi.kiev.ua
- 4. MARKET PREDICTION (MODELS MODIFICATION) Four model types for stock market prediction: • multiplicative form of threshold function and constant broker’s reputation matrix • multiplicative form of threshold function and variable broker’s reputation matrix • additive form of threshold function and constant broker’s reputation matrix • additive form of threshold function and variable broker’s reputation matrix
- 5. MARKET PREDICTION (ADVANTAGES) NEW! Advantages Predicting power Accounting the collection of brokers behavior Concrete results on brokers interaction in dynamics Can easy operate with many brokers situation Functional dependences which adequate to real processes of broker interaction
- 6. MARKET PREDICTION (INPUT PARAMETYERS) • Brokers number at stock market; • Broker’s reputation matrix; • Broker’s financial resources ; • Influence of broker’s intention on request; • Initial requests vector; • The training velocity and forgetting velocity (for case of variable reputation matrix); • Prediction horizon.
- 7. MARKET PREDICTION (INPUT DATA FORMATS) Dynamics of stock course 30 25 max min Closed 20 15
- 8. MARKET PREDICTION (MODELING RESULTS) • Broker’s request matrix for all time of negotiation • Reputation matrix with changes (for model with variable reputation matrix) • Picture of requests dynamics at the time interval of negotiation • Picture of the ‘network energy’
- 9. MARKET PREDICTION (SOFTWARE INTERFACE 1) Multiplicative model with variable reputation matrix
- 11. MARKET PREDICTION (SOFTWARE INTERFACE 3) Multiplicative model with constant reputation matrix