Deciphering Customer Satisfaction: A Machine Learning-Oriented Method Using Agglomerative Clustering for Predictive Modeling and Feature Selection

特征选择 层次聚类 选择(遗传算法) 特征(语言学) 聚类分析 机器学习 顾客满意度 计算机科学 人工智能 企业管理 数据挖掘 工程类 业务 营销 哲学 工商管理 语言学
作者
Nisrine Rezki,Mohamed Mansouri,Rachid Oucheikh
出处
期刊:Management Systems in Production Engineering [P.A. NOVA S.A.]
卷期号:33 (1): 60-70
标识
DOI:10.2478/mspe-2025-0007
摘要

Abstract In contemporary enterprises, customer satisfaction analysis has become a critical area of concentration. Being able to understand and predict customer satisfaction is becoming more and more important as companies try to develop and launch new products. Leveraging customer data intelligently and employing robust data analytics techniques are essential for meeting this imperative. With this objective in mind, the study proposes a machine learning-based approach to analyze and discern the variables influencing customer satisfaction. Specifically, the study utilizes agglomerative clustering for data segmentation and feature identification, followed by a Random Forest Classifier as machine learning (ML) model for prediction. Performance metrics such as accuracy, recall, precision and F1-score are employed for model evaluation, ensuring robustness and reliability in the predictive process. Furthermore, it aims to predict the impact of enhancing specific product attributes on customer satisfaction. To provide a tangible demonstration of the proposed methodology, a comprehensive case study is conducted. By systematically integrating clustering techniques into the feature selection and modeling process, this framework furnishes a structured methodology for data-driven decision-making and predictive analytics. This holistic approach not only enriches the comprehension of intricate datasets but also facilitates the development of resilient predictive models characterized by enhanced accuracy and interpretability. By segmenting customers based on their responses, we discerned specific areas of satisfaction and dissatisfaction, providing actionable insights for targeted strategies aimed at improving overall satisfaction. The insights and customer clustering derived from this study can guide these targeted strategies to enhance customer satisfaction and inform future product development initiatives.

科研通智能强力驱动
Strongly Powered by AbleSci AI
科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
celinewu完成签到,获得积分10
1秒前
1秒前
格子大王完成签到,获得积分10
2秒前
默默zzz完成签到 ,获得积分10
2秒前
2秒前
Dautless完成签到,获得积分10
3秒前
万能图书馆应助彳亍采纳,获得10
3秒前
3秒前
5秒前
刘威完成签到,获得积分10
6秒前
TL发布了新的文献求助30
6秒前
量子星尘发布了新的文献求助10
6秒前
wanchao完成签到,获得积分10
8秒前
9秒前
科研通AI6.2应助海苔噗噗采纳,获得10
9秒前
雨碎寒江发布了新的文献求助10
9秒前
10秒前
CodeCraft应助11采纳,获得30
10秒前
10秒前
11秒前
妮儿发布了新的文献求助10
11秒前
Akim应助ltr采纳,获得10
11秒前
显眼包完成签到,获得积分20
11秒前
zzzzzz发布了新的文献求助10
11秒前
12秒前
han222发布了新的文献求助10
12秒前
炎之域完成签到,获得积分10
12秒前
善学以致用应助Lexi采纳,获得10
13秒前
14秒前
英姑应助TE采纳,获得10
14秒前
搜集达人应助zzzzz采纳,获得10
15秒前
Jasper应助Bio采纳,获得10
15秒前
小桔青山发布了新的文献求助10
15秒前
檬檬完成签到,获得积分10
16秒前
16秒前
甘橘发布了新的文献求助10
19秒前
你好完成签到,获得积分10
19秒前
19秒前
19秒前
活力忆雪发布了新的文献求助10
19秒前
高分求助中
(应助此贴封号)【重要!!请各用户(尤其是新用户)详细阅读】【科研通的精品贴汇总】 10000
Kinesiophobia : a new view of chronic pain behavior 2000
Cronologia da história de Macau 1600
Earth System Geophysics 1000
Bioseparations Science and Engineering Third Edition 1000
Lloyd's Register of Shipping's Approach to the Control of Incidents of Brittle Fracture in Ship Structures 1000
BRITTLE FRACTURE IN WELDED SHIPS 1000
热门求助领域 (近24小时)
化学 材料科学 医学 生物 工程类 纳米技术 有机化学 物理 生物化学 化学工程 计算机科学 复合材料 内科学 催化作用 光电子学 物理化学 电极 冶金 遗传学 细胞生物学
热门帖子
关注 科研通微信公众号,转发送积分 6127570
求助须知:如何正确求助?哪些是违规求助? 7955220
关于积分的说明 16507063
捐赠科研通 5246496
什么是DOI,文献DOI怎么找? 2802122
邀请新用户注册赠送积分活动 1783379
关于科研通互助平台的介绍 1654490