[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-123504-en":3,"doc-seo-123504-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},123504,1099514068365,"Aurelia","https://ap-avatar.wpscdn.com/avatar/10000253d8d9f28188e?_k=1776742907772140068",8,"Research & Report","Applications of Panel Data and Machine Learning Algorithms for Data-Driven Decision-Making - Doctoral Thesis","This doctoral dissertation advances empirical decision-making using panel data combined with machine learning methods. The research studies how personality traits relate to socio-economic and business outcomes, testing separability of age, period, and cohort effects and their lifecycle evolution. It compares recurrent neural networks and transformers for customer churn prediction using large anonymized panel data, and proposes an RNN-plus-extreme gradient boosting strategy. It further estimates and evaluates demand price elasticities with machine learning and unbalanced panels in e-commerce through bias-variance comparisons and multiple pricing experiments.","Applications of Panel Data and Machine Learning Algorithms for Data-Driven Decision-Making  \nDOCTORAL THESIS  \nto acquire the academic degree of  \ndoctor rerum politicarum  \n(Doctor of Economics and Management Science)  \nsubmitted to the  \nSchool of Business and Economics of  \nHumboldt-Universität zu Berlin  \nby  \nM.Sc. Christian Gary Mena Lecoña  \nPresident of Humboldt-Universität zu Berlin: Prof. Dr. Julia von Blumenthal  \nDean of the School of Business and Economics: Prof. Dr. Daniel Klapper  \nReviewers: 1. Prof. Dr. Stefan Lessmann  \n2. Prof. Dr. Daniel Guhl  \nDate of Colloquium: July 9, 2025  \nAbstract  \nThis dissertation focuses on empirical applications of panel data and machine learning algorithms for better decision-making. The first chapter uses data from the German Socio-Economic Panel to study personality traits and the hypothesis of separability of age, period, and cohort effects. There is mounting evidence in the literature on the importance of various personality traits to predict socio-economic and business outcomes. This motivates us to investigate their life cycle profile after controlling for cohort and period effects. Our results from a battery of specification tests provide evidence for theseparability of these effects for a number of the traits. Our analysis also reveals that they follow a stable, but not constant, evolution over the life cycle.  \nThe second chapter evaluates the performance gains that recurrent neural networksand transformers bring to customer churn prediction. The empirical evaluation is based on anonymized panel data from a European provider of financial services that contains data of 480 thousand customers over a period of 48 months. The results show that recurrent neural networks, in our application, perform better than transformer-based architecturesas measured by top-decile lift, area under the receiver operating curve, and the expected maximum profit criterion. When the data includes a mix of time-varying and static features, the metrics suggest that it is preferable to process the latter with a feed-forward neural network and then concatenate the outputs with the data representations of the recurrent network. Finally, the experiments reveal that using the output of the recurrent neural network as input of an extreme gradient boosting classifier offers the best out-of-sample performance among the considered models.  \nThe third and fourth chapters focus on estimating and evaluating demand price elasticities based on machine learning algorithms and unbalanced panel data for applications in e-commerce. The third chapter starts by comparing the bias and variance of double selection estimators against variations of a dynamic panel double machine learning model. Controlled simulations show that the latter has a lower bias but larger variance. We use this insight to test if the estimated elasticities from the least biased version of the double machine learning model allow us to identify price-sensitive products. The hypothesis is that for product groups with an estimated elasticity above one in absolute value, i.e. , price-sensitive, a randomly assigned price reduction of ten percent leads to higher revenue compared to maintaining the price constant. We collaborated with a large e-retailer and conducted two pricing experiments showing that this pricing policy leads to an average increase in revenue and profit. Thus, our experimental evidence provides support in favor of the double machine learning estimates and their application in pricing decision-making. The fourth chapter documents two new pricing experiments that improve in two dimensions the evaluation of the double machine learning estimates. First, we implement a cluster-based randomization that allows us to assess the effect of pricing spillovers.  \nSecond, we expand the evaluation to test the economic value of the point estimates to identify elastic and inelastic products. The results show that, when we only consider price reduc","cbCaiq4bR64499V5","https://ap.wps.com/l/cbCaiq4bR64499V5","pdf",7174042,1,140,"English","en",105,"# Abstract\n## Personality traits and age-period-cohort separability\n## Customer churn prediction with RNNs and transformers\n## Demand price elasticities in e-commerce\n## Pricing experiments and policy evaluation","[{\"question\":\"What does the dissertation investigate in its first chapter?\",\"answer\":\"It uses German Socio-Economic Panel data to study personality traits and test whether age, period, and cohort effects are separable. It also examines how these traits evolve over the life cycle after controlling for cohort and period effects.\"},{\"question\":\"How does the dissertation evaluate customer churn prediction models?\",\"answer\":\"It compares recurrent neural networks and transformer-based architectures on anonymized European financial-services panel data. The results favor RNNs by several metrics, and the best out-of-sample performance comes from feeding RNN outputs into an extreme gradient boosting classifier.\"},{\"question\":\"How are demand price elasticities applied and tested for e-commerce pricing decisions?\",\"answer\":\"The dissertation estimates price elasticities using machine learning with unbalanced panel data, then evaluates whether the least-biased double machine learning estimates can identify price-sensitive products. It validates the approach through pricing experiments, including cluster-based randomization for spillover effects and tests of economic value for elastic versus inelastic products.\"}]","Applications of Panel Data and Machine Learning Algorithms for Data-Driven Decision-Making - Doctoral Thesis | PDF",1785816903,353,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"applications-of-panel-data-and-machine-learning-algorithms-for-data-driven-decision-making-doctoral-thesis","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/applications-of-panel-data-and-machine-learning-algorithms-for-data-driven-decision-making-doctoral-thesis/123504/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-04",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What does the dissertation investigate in its first chapter?","Question",{"text":75,"@type":76},"It uses German Socio-Economic Panel data to study personality traits and test whether age, period, and cohort effects are separable. It also examines how these traits evolve over the life cycle after controlling for cohort and period effects.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does the dissertation evaluate customer churn prediction models?",{"text":80,"@type":76},"It compares recurrent neural networks and transformer-based architectures on anonymized European financial-services panel data. The results favor RNNs by several metrics, and the best out-of-sample performance comes from feeding RNN outputs into an extreme gradient boosting classifier.",{"name":82,"@type":73,"acceptedAnswer":83},"How are demand price elasticities applied and tested for e-commerce pricing decisions?",{"text":84,"@type":76},"The dissertation estimates price elasticities using machine learning with unbalanced panel data, then evaluates whether the least-biased double machine learning estimates can identify price-sensitive products. It validates the approach through pricing experiments, including cluster-based randomization for spillover effects and tests of economic value for elastic versus inelastic products.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]