應(yīng)理學(xué)院邀請(qǐng)，英國(guó)南安普頓大學(xué)Houduo Qi(戚厚鐸)教授來(lái)我校進(jìn)行學(xué)術(shù)交流，并做學(xué)術(shù)報(bào)告。
報(bào)告時(shí)間：2016年12月5日周一下午2:00—3:00
報(bào)告地點(diǎn)：西教五416（理學(xué)院）
報(bào)告題目：Can You Trust your Correlation Matrix -- a question from finance
報(bào)告摘要：
The concept of correlations plays a very important role in risk management in finance and a correlation matrix has to satisfy a set of mathematical properties. However, when the historical data is incomplete due to various market conditions, the sampling correlation matrix often violates those properties. Worse than such a situation is the scenario when correlations have to be humanly set to fence off risks from certain disastrous events (this is also known as correlation stress testing in finance). Therefore, adjusting an invalid correlation matrix to one that is to satisfy those mathematical properties is a problem of paramount importance. An equally important requirement from practice is the speed of such an adjustment. We show that this can be done through optimization. We introduce a fast algorithm that exploits the second-order sparsity to make a linear equation solver fast enough to meet the practical need. The resulting algorithm has now been widely used in finance industry (Based on joint work with Defeng Sun).
報(bào)告人簡(jiǎn)介：戚厚鐸，1990年本科畢業(yè)于北京大學(xué)概率統(tǒng)計(jì)系，1996年于中國(guó)科學(xué)院應(yīng)用數(shù)學(xué)研究所獲理學(xué)博士學(xué)位，導(dǎo)師韓繼業(yè)研究員。博士畢業(yè)后先后在中國(guó)科學(xué)院計(jì)算數(shù)學(xué)與科學(xué)工程計(jì)算研究所和澳大利亞新南威爾士大學(xué)從事博士后研究工作，2004年起任職于英國(guó)南安普頓大學(xué)?，F(xiàn)任國(guó)際SCI期刊Asia-Pacific Journal of Operational Research的Area Editor (Optimization)以及Mathematical Programming Computation的Associate Editor.
