卡特三农学术论坛之246期(大数据与人工智能专题首讲)通知 主讲人:于晓华-尊龙官方平台

卡特三农学术论坛之246期(大数据与人工智能专题首讲)通知 主讲人:于晓华

编辑:浙江大学中国农村发展研究院 作者: 时间:2023-09-15 访问次数:31


卡特三农学术论坛之246期(大数据与人工智能专题首讲)

暨“食物系统与政策”海外大师联合工作室(fsp lab)系列讲座第七讲

 “求是智库”系列讲座资源环境与农业发展大讲堂第20

于晓华:heterogeneity and dynamics of food price inflations in oecd countries: evidence from machine learning


讲座时间2023922日(周五)9:30-12:30

讲座地点:浙江大学紫金港校区公共管理学院112报告厅

本期主题heterogeneity and dynamics of food price inflations in oecd countries: evidence from machine learning

主讲嘉宾

于晓华,德国哥廷根大学发展与转型国家农业经济学讲席教授,农业经济与农村发展系系主任。2009年毕业于美国宾州州立大学,获得农业经济学与人口学双博士。2009年起任教于德国哥廷根大学,主要研究方向为农业经济学、发展经济学、以及行为经济学等。20229月至20233月在康奈尔大学johnson商学院担任访问教授。曾经或现在担任agricultural economicschina economic review等国际期刊副主编或者编委;以及非洲农业研究论坛forum for agricultural research in africa、联合国粮农组织2024年度报告,以及德国转型国家农业经济研究所(iamo)等国际组织的科学咨询委员等职务。在american journal of agricultural economics, journal of economic dynamics and control, land economics, food policy, land use policy, journal of rural studies 等期刊发表论文多篇。

内容摘要since 2020, food prices are on the rise, resulting in an ongoing global food price crisis. rising food prices not only represent a monetary burden for oecd countries but are directly linked to increased costs of living and food security. the precise prediction of food prices through relevant features consequently represents a key instrument to the design of responsive policy measures that effectively address rises in food prices and protect vulnerable or food-insecure households, both in oecd countries and net food importing countries. through dynamic time warping (dtw) time-series clustering, we aim to capture long-term divergences in food cpi between member countries. cluster-specific predictors of food prices are identified through long-short term memory (lstn) neural networks. we show that long-term food price trends are cluster-specific and do not follow a uniform pattern within the oecd. food prices between clusters are determined by a differential set of predictors. the marginal contribution of singular predictors changes over time, highlighting the volatile and transient character of predictors. this implies that there is no singular set of predictors that continuously shapes food prices but rather a diverse set of predictors that interacts with time. the results highlight the complex and time-contingent character of food price mechanisms that should be taken into consideration when designing price policies that aim to address surges in food prices in the future.

关键词:oecd, long-term food prices, time series clustering (dtw), long-short term memory (lstm) neural networks

主持人:茅锐,浙江大学中国农村发展研究院副院长,农业经济与管理系主任。

网站地图