| 比賀江 文子 | D, 中間発表 | 数理情報学 | 池田 和司, | 笠原 正治, | 吉本 潤一郎, | 久保 孝富, | 福嶋 誠, | 日永田 智絵 |
| BIKERI ADLINE KERUBO | D, 中間発表 | 数理情報学 | 池田 和司, | 笠原 正治, | 吉本 潤一郎, | 福嶋 誠, | 日永田 智絵 | |
|
title: A Hawke’s model approach to modeling price spikes in electricity markets
abstract: The Japanese Electric Power eXchange (JEPX) provides a platform for trading of electric energy in a manner similar to more traditional financial markets. As the number of market agents increase there is an increasing need for effective price forecasting models. Electricity price data is observed to exhibit periods of relatively stable i.e., low-magnitude, low-variance prices interspersed by periods of higher prices accompanied by larger uncertainty. The price data time series therefore exhibits a temporal non-stationarity characteristic that is difficult to capture with typical time-series modeling frameworks. We therefore propose modeling the electricity price generating process as consisting of two sources of noise: 1) a low-variance IID noise that explains small inter-day variations and, 2) relatively larger magnitude disturbances that explain periodic spikes in prices. We define spikes as observing prices above a predefined threshold and model these as a marked Hawkes process whereby the occurrence of a spike event increases the probability of observing more spikes in the period immediately following the event. Here, we present preliminary results detailing Bayesian estimates of the model parameters when individual trading time slots are considered as univariate processes. In the next step of the research, we intend to extend to a multivariate model and evaluate the effectiveness of the model in forecasting spikes. Finally, we will develop a method for simultaneously estimating both the internal factors causing the low-magnitude variations and the external factors causing the large magnitude spikes. language of the presentation: English | ||||||||
| ZAVIALOV IGOR | D, 中間発表 | 数理情報学 | 池田 和司, | 笠原 正治, | 吉本 潤一郎, | 福嶋 誠, | 日永田 智絵 | |
|
title: Systemic risk and stability of a financial system
abstract: Systemic risk in a financial system is determined by the possibility of the "domino effect" when the failure of one participant leads to the failure of others and the system’s collapse. Rare extreme events such as financial crises happen several times in a century and are characterized by higher systemic risk. The goal of this research is to estimate or classify systemic risk of the financial system using modern Machine Learning approaches, analyze the stability of a financial system using advances in chaos theory and other methods. Although the intuition behind the term of systemic risk is clear, the precise mathematical definition is difficult to formulate because of the system complexity and interconnectivity. In this presentation, I will discuss possible directions for systemic risk analysis. As a particular example, I will consider application of Kron-Kerbosh algorithm to the problem of finding SIFI (significantly important financial institutions). language of the presentation: English | ||||||||