國立虎尾科技大學 |

Structural Neural Networks Meet Piecewise Exponential Models on Customer Acquisition and Retention.

Record Type:	Language materials, manuscript : Monograph/item
Title/Author:	Structural Neural Networks Meet Piecewise Exponential Models on Customer Acquisition and Retention./
Author:	Cai, Chuan.
Description:	1 online resource (150 pages)
Notes:	Source: Dissertations Abstracts International, Volume: 85-08, Section: B.
Contained By:	Dissertations Abstracts International85-08B.
Subject:	Higher education administration. -
Online resource:	click for full text (PQDT)
ISBN:	9798381721775

Structural Neural Networks Meet Piecewise Exponential Models on Customer Acquisition and Retention.
Cai, Chuan.

Structural Neural Networks Meet Piecewise Exponential Models on Customer Acquisition and Retention. - 1 online resource (150 pages)

Source: Dissertations Abstracts International, Volume: 85-08, Section: B.

Thesis (Ph.D.)--University of Delaware, 2024.

Includes bibliographical references

This dissertation examines the dynamics of customer acquisition and retention within the context of higher education with students being recognized as customers. Given the substantial reliance of higher education institutions on tuition fees for operating revenue, student recruitment and retention emerge as pivotal aspects of enrollment management. Furthermore, retention and timely graduation are used to measure institutional reputation and accountability. This study pursues three main objectives: understanding applicant deposit decisions during the admission process, predicting matriculated students' dropout risks, and exploring the impact of student loan debt on timely graduation. The first segment analyzes the determinants of deposit decisions of out-of-state students admitted to the University of Delaware across three academic years. Utilizing three Bayesian hierarchical piecewise exponential models, we deduce that factors like gender and recruitment events exhibit time-varying effects, while others like financial aid remain stable within an academic year but vary across years. The baseline desire to deposit intensifies as deadlines near, though this trajectory shifts annually. Insights derived inform the Admissions Office's marketing and recruitment tactics. The second segment introduces a hybrid model, merging a structural neural network with a piecewise exponential model, to predict college attrition. Benchmarking against two alternative models, the hybrid model demonstrates superior or comparable predictive prowess for the University of Delaware across three springs. Categorizing predictors into academic, economic, and socio-demographic facets reveals academic indicators as key discriminants between students who drop out and those retained, especially from freshman to junior years. Emphasis on academic assessments in intervention strategies is thus recommended.The third segment evaluates the impact of student loan debt on six-year graduation rates by department, over a span of five years. Leveraging five Bayesian hierarchical models, the findings illustrate a pronounced department-wise loan debt effect on first-year students, which attenuates as they advance academically. Tailored financial aid policies, considering academic departments, are posited to amplify the efficient utilization of institutional financial resources. For universities mulling over department-specific financial aid policies, initiation with randomized trials for first-year students is advised.In summary, this dissertation introduces innovative strategies for strategic enrollment management, encompassing admission, retention, and graduation considerations. Particular attention is given to the dynamic nature of applicant deposit decisions, the development of predictive models for student attrition, and the department-specific effects of student loan debt on graduation rates.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2024

Mode of access: World Wide Web

ISBN: 9798381721775Subjects--Topical Terms:

1148709
Higher education administration.
Subjects--Index Terms:

Customer acquisitionIndex Terms--Genre/Form:

554714
Electronic books.

Structural Neural Networks Meet Piecewise Exponential Models on Customer Acquisition and Retention.
LDR:04321ntm a22003977 4500 001 1146247
005 20240812064351.5
006 m o d
007 cr bn ---uuuuu
008 250605s2024 xx obm 000 0 eng d
020 $a 9798381721775
035 $a (MiAaPQ)AAI30819427
035 $a AAI30819427
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Cai, Chuan. $3 1471605
245 1 0 $a Structural Neural Networks Meet Piecewise Exponential Models on Customer Acquisition and Retention.
264 0 $c 2024
300 $a 1 online resource (150 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 85-08, Section: B.
500 $a Advisor: Fleischhacker, Adam J.
502 $a Thesis (Ph.D.)--University of Delaware, 2024.
504 $a Includes bibliographical references
520 $a This dissertation examines the dynamics of customer acquisition and retention within the context of higher education with students being recognized as customers. Given the substantial reliance of higher education institutions on tuition fees for operating revenue, student recruitment and retention emerge as pivotal aspects of enrollment management. Furthermore, retention and timely graduation are used to measure institutional reputation and accountability. This study pursues three main objectives: understanding applicant deposit decisions during the admission process, predicting matriculated students' dropout risks, and exploring the impact of student loan debt on timely graduation. The first segment analyzes the determinants of deposit decisions of out-of-state students admitted to the University of Delaware across three academic years. Utilizing three Bayesian hierarchical piecewise exponential models, we deduce that factors like gender and recruitment events exhibit time-varying effects, while others like financial aid remain stable within an academic year but vary across years. The baseline desire to deposit intensifies as deadlines near, though this trajectory shifts annually. Insights derived inform the Admissions Office's marketing and recruitment tactics. The second segment introduces a hybrid model, merging a structural neural network with a piecewise exponential model, to predict college attrition. Benchmarking against two alternative models, the hybrid model demonstrates superior or comparable predictive prowess for the University of Delaware across three springs. Categorizing predictors into academic, economic, and socio-demographic facets reveals academic indicators as key discriminants between students who drop out and those retained, especially from freshman to junior years. Emphasis on academic assessments in intervention strategies is thus recommended.The third segment evaluates the impact of student loan debt on six-year graduation rates by department, over a span of five years. Leveraging five Bayesian hierarchical models, the findings illustrate a pronounced department-wise loan debt effect on first-year students, which attenuates as they advance academically. Tailored financial aid policies, considering academic departments, are posited to amplify the efficient utilization of institutional financial resources. For universities mulling over department-specific financial aid policies, initiation with randomized trials for first-year students is advised.In summary, this dissertation introduces innovative strategies for strategic enrollment management, encompassing admission, retention, and graduation considerations. Particular attention is given to the dynamic nature of applicant deposit decisions, the development of predictive models for student attrition, and the department-specific effects of student loan debt on graduation rates.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2024
538 $a Mode of access: World Wide Web
650 4 $a Higher education administration. $3 1148709
650 4 $a Statistics. $3 556824
650 4 $a Higher education. $3 1148448
653 $a Customer acquisition
653 $a Customer retention
653 $a Piecewise exponential model
653 $a Structural neural network
655 7 $a Electronic books. $2 local $3 554714
690 $a 0446
690 $a 0463
690 $a 0310
690 $a 0745
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a University of Delaware. $b Financial Services Analytics. $3 1466246
773 0 $t Dissertations Abstracts International $g 85-08B.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30819427 $z click for full text (PQDT)