ترسیم و تفسیر درخت تصمیم‌گیری مشتریان بازارچه مرزی اینچه برون با استفاده از رویکرد ترکیبی

یکی از بازیگران اصلی بازارچه‌های مرزی که موجودیت و رونق این بازارچه‌ها به حضور آن‌ها گره خورده است مشتریان هستند که رفتار خرید آن‌ها در سایه‌ی سایر ابعاد مهم بازارچه‌های مرزی، کمتر مورد توجه قرار گرفته است. هدف پژوهش حاضر، ترسیم و تشریح درخت تصمیم‌گیری مشتریان بازارچه مرزی اینچه برون با استفاده از رویکرد ترکیبی می‌باشد، بنابراین این پژوهش کاربردی بوده، از حیث هدف توصیفی پیمایشی و در قالب مطالعات مقطعی است که در دو گام مرتبط با هم صورت پذیرفت. جامعه آماری پژوهش در بخش کیفی شامل مشتریان بازارچه بودند که از طریق روش نمونه‌گیری نظری درمجموع 21 مصاحبه انجام شد. جامعه آماری پژوهش در بخش کمی مشتریان بازارچه بودند که 389 پرسشنامه پس از پاک‌سازی و غربال اولیه قابل‌استفاده تشخیص داده شدند که تحلیل‌ها روی آن‌ها صورت پذیرفت. در گام اول با استفاده از روش تحلیل محتوا، 34 شاخص تاثیرگذار بر خرید مشتریان استخراج‌شده  و بر مبنای آن، پرسشنامه‌ی پژوهش طراحی شد که پایایی و روایی آن به ترتیب با استفاده از آلفای کرونباخ و تحلیل عاملی تاییدی بررسی و تایید شد. در گام دوم از رویکردهای کمی (خوشه‌بندی و درخت تصمیم) جهت بررسی جامعه آماری استفاده شد. به‌منظور تشریح جامعه پژوهش، 389 مشتری بازارچه مرزی اینچه برون با استفاده از شاخص دیویس بولدین در سه خوشه مجزا از هم خوشه‌بندی و مورد بررسی قرار گرفتند. به‌منظور ترسیم الگوی تصمیم‌گیری مشتریان در خرید به بازارچه از درخت تصمیم استفاده شد. به‌کارگیری درخت تصمیم نشان داد که با استفاده از شاخص‌های تاثیرگذار بر رفتار خرید مشتریان می‌توان نرخ تکرار خرید آن‌ها در سه خوشه اول، دوم و سوم را به ترتیب با دقت 0.851، 0.826 و 0.852 پیش‌بینی کرد. نتایج ترسیم و تفسیر درخت‌های تصمیم بهینه مشتریان به تفکیک تحلیل خوشه‌ای نشان داد در خوشه اول شاخص سرویش بهداشتی، سطل زباله، آب آشامیدنی، در خوشه دوم شاخص قیمت محصولات و تخفیفات و در خوشه سوم شاخص امکانات روشنایی جاده نقش به سزایی در تکرار خرید مشتری داشته و گره صفر درخت تصمیم هر خوشه را به خود اختصاص داده‌اند.

drawing and interpreting the decision tree of the border market inche barun customers using a hybrid approach

aim and introduction: one of the key players in border bazaars, whose existence and prosperity are closely linked to their presence, is the customer. however, customer buying behavior has often been overlooked in favor of other significant aspects of border bazaars. this research employs the customer journey map theory to identify the factors influencing the shopping journey of customers. customer journey maps encompass both online and offline interactions, as well as pre and post purchase encounters. they are valuable tools for gaining insights into how a series of experiences unfolds over time through sequential, service based processes. the growing focus on customer journeys reflects management’s interest in enhancing value by improving customer experiences.methodology: the purpose of the current research is to develop a decision making model for customers of the inche barun border market using a hybrid approach. this study is applied in nature, employing a descriptive survey methodology and conducted as a cross sectional study, which was carried out in two related phases. the statistical population for the qualitative aspect of the research comprised visitors and customers of the border market. the study employed a theoretical sampling method, continuing until theoretical saturation was achieved, resulting in a total of 21 interviews. the statistical population for this research consisted of visitors to the border market. after cleaning and preliminary screening, 389 questionnaires were deemed usable for analysis. in the first step, we employed content analysis to identify 34 indicators that influence the purchasing journey of customers. based on these indicators, a research questionnaire was developed, and its reliability and validity were confirmed using cronbach’s alpha and confirmatory factor analysis, respectively. in the second step, quantitative methods, including clustering and decision tree analysis, were utilized to examine the statistical population. in order to describe the research population, 389 visitors of the inche barun market were grouped and analyzed into three separate clusters using the davis bouldin index. a decision tree was employed to illustrate the pattern of customers’ decision making during their shopping trips to the bazaar.finding: the findings from the qualitative stage of the research resulted in the identification of 34 concepts. the classification of these concepts based on their semantic and conceptual affinities led to the establishment of six main categories. since these concepts and categories reflect the desires and expectations of customers in their own language, they will be highly beneficial for officials and decision makers in this field. the application of decision trees demonstrated that by utilizing indicators that influence customer buying behavior, it is possible to predict their return rates in the first, second, and third clusters with accuracies of 0.851, 0.826, and 0.852, respectively. this indicates that the decision making models among the three clusters do not significantly differ in terms of accuracy. however, the analysis of customer decision making patterns reveals that the criteria for repeating shopping trips vary considerably among the clusters. specifically, in the first, second, and third clusters, the zero nodes correspond to the following indices: sanitary ware, trash cans, drinking water, product prices and discounts, and road lighting facilities.discussion and coclusion: the results of the research on cluster analysis indicated that customers interact with various contact points during their shopping trips to border markets. based on these interactions, distinct clusters can be identified, demonstrating that the statistical population under investigation is heterogeneous. additionally, the application of the decision tree model to analyze decision patterns revealed that customer buying behavior is complex, hierarchical, and multi criteria. the clustering of border market customers reveals that, although shoppers engage with various and diverse contact points throughout their shopping journey, the significance of these points varies in both level and extent among different travelers. many of these points do not play a central role in the decision making process, which differentiates the sample. statistics grouped into three clusters confirm this issue. the analysis of the statistical sample from the current research, which is based on 34 indicators influencing the customer’s purchase journey, reveals that the target population in the cross border market is not homogeneous. presenting a single, uniform program will not adequately address the diverse tastes and interests of all consumers. in other words, while a standardized approach may effectively estimate the demands of one cluster, it could yield entirely different results for another cluster. the results of drawing and interpreting the optimal decision tree for customers through cluster analysis revealed that, in the first cluster, the indices of sanitary napkins, trash cans, and drinking water were significant factors influencing repeat customer purchases. in the second cluster, the indices of product prices and discounts played a crucial role, while in the third cluster, the indices of road lighting facilities were important. the zero node of the decision tree for each cluster is assigned to these factors.