دانلود مقاله ISI انگلیسی شماره 22201
عنوان فارسی مقاله

تسهیل متقابل فروش در بازار مخابرات تلفن همراه به منظور توسعه مدل طبقه بندی مشتریان بر اساس تکنیک های داده کاوی ترکیبی

کد مقاله سال انتشار مقاله انگلیسی ترجمه فارسی تعداد کلمات
22201 2011 8 صفحه PDF سفارش دهید 6200 کلمه
خرید مقاله
پس از پرداخت، فوراً می توانید مقاله را دانلود فرمایید.
عنوان انگلیسی
Facilitating cross-selling in a mobile telecom market to develop customer classification model based on hybrid data mining techniques
منبع

Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)

Journal : Expert Systems with Applications, Volume 38, Issue 5, May 2011, Pages 5005–5012

کلمات کلیدی
طبقه بندی مشتریان - الگوریتم های ژنتیکی - رگرسیون لجستیک - شبکه های عصبی مصنوعی - درخت های تصمیم گیری - بازار مخابرات تلفن همراه
پیش نمایش مقاله
پیش نمایش مقاله تسهیل متقابل فروش در بازار مخابرات تلفن همراه به منظور توسعه مدل طبقه بندی مشتریان بر اساس تکنیک های داده کاوی ترکیبی

چکیده انگلیسی

As the competition between mobile telecom operators becomes severe, it becomes critical for operators to diversify their business areas. Especially, the mobile operators are turning from traditional voice communication to mobile value-added services (VAS), which are new services to generate more average revenue per user (ARPU). That is, cross-selling is critical for mobile telecom operators to expand their revenues and profits. In this study, we propose a customer classification model, which may be used for facilitating cross-selling in a mobile telecom market. Our model uses the cumulated data on the existing customers including their demographic data and the patterns for using old products or services to find new products and services with high sales potential. The various data mining techniques are applied to our proposed model in two steps. In the first step, several classification techniques such as logistic regression, artificial neural networks, and decision trees are applied independently to predict the purchase of new products, and each model produces the results of their prediction as a form of probabilities. In the second step, our model compromises all these probabilities by using genetic algorithm (GA), and makes the final decision for a target customer whether he or she would purchase a new product. To validate the usefulness of our model, we applied it to a real-world mobile telecom company’s case in Korea. As a result, we found that our model produced high-quality information for cross-selling, and that GA in the second step contributed to significantly improve the performance

مقدمه انگلیسی

Along with the rapid growth of mobile telecom market in recent years, the mobile telecom markets in the world are becoming saturated. As traditional voice communication services have become widespread, mobile telecom operators have experienced difficulties in attracting more customers, which has led to fierce competition amongst the mobile operators. To overcome the current crisis, the mobile operators are focusing more on the value-added services (VAS) 1 market such as communication, entertainment, transaction, and information services, instead of the ordinary voice communications market since the VAS market returns higher average revenue per user (ARPU). In other words, the mobile operators are trying to cross-sell their new mobile telecom service ( Kuo & Chen, 2006). However, it is never easy for the operators to make their subscribers to use mobile VAS. There are two types of the subscribers who do not buy or use mobile VAS. The first type of subscribers are the people who have no interests in using mobile VAS or are not able to use it. They simply think that mobile VAS do not offer any benefits. The second type is the subscribers who may have interests and may find benefits in using mobile VAS, but are not aware of the availability of the services. These second type subscribers offer great business opportunities for mobile service operators. Thus, it becomes critical for the operators to find the appropriate prospects for using mobile VAS. Recently, most mobile telecom operators utilize CRM (customer relationship management) systems in order to cumulate various types of data from their subscribers, such as demographic information and the patterns for using voice communication, in their DB (database) or DW (data warehouse). These data can be useful to find the appropriate prospects for using mobile VAS since they may be used as the cues for understanding customers’ life styles and value system. As a result, the effective application of the cumulated data on the existing customers and their usage pattern for old products or services may serve as a core competence of companies including mobile telecom operators. To meet these kinds of needs, we propose an innovative customer classification model, which may be used to facilitate cross-selling. This model can predict whether a customer would purchase new products or services by using the cumulated data on the customer. Our model consists of two steps, which applies several data mining techniques. The first step builds multiple classification models simultaneously by using data mining techniques for classification such as logistic regression, artificial neural networks, and decision trees. Each of these models produces the probability of a customer to purchase or use new products. In the second step, all of these probabilities are compromised by genetic algorithms (GA), and final scores for customers are calculated, which can be used as a guide to find prospects for cross-selling new products. Furthermore, the thresholds that are used to interpret the final score to the final decision are also optimized by GA in this step. Though the customer classification model proposed in this paper can be applied to various kinds of industries or markets, we apply it to an existing mobile telecom company’s case in Korea as cross-selling is crucial for competence and survival in the mobile telecom industry. We organize this article as follows. In Section 2, we briefly review the theoretical background, and propose our research model – a new customer classification model – in the next section. Section 4 presents the research design and experiments. The empirical results are presented and discussed in Section 5. The final section suggests the contributions and limitations of this study.

نتیجه گیری انگلیسی

In this paper, we have suggested a novel classification model, which combines several heterogeneous classifiers in an effective way. Our model combines the prediction results of each classifier by weighted-averaging method, although conventional combination methods just combines the prediction results by applying simple techniques such as averaging and majority voting. Also, our proposed model uses two thresholds (i.e. upper and lower thresholds) in order to interpret and integrate the continuous scores into final decision more effectively. As a result, our model may not yield explicit results for cases that display vague patterns. Nevertheless, it can boost the prediction accuracy dramatically. To optimize the different factors simultaneously, we use GA as a tool to optimize the combination weights and the two thresholds. From the results of the experiment, we show that our proposed model can improve the prediction accuracy in the case of customer classification, though the coverage ratio decreases more or less. The proposed model in this study seems to be very useful in some areas in which require high prediction accuracy for the limited cases. The cross-selling application suggested in this paper may be a good example. In reality, our model can be applied to facilitate cross-selling in a mobile telecom market, as presented in Fig. 5. By sending advertising SMS or MMS to the target customers carefully selected by our model, cross-selling can be facilitated. Meanwhile, direct and indirect costs can be saved by avoiding sending messages to inappropriate customers.Our model can be applied in other marketing or CRM domains where finding prospects to purchase a new product or service is required. In addition, it may contribute to some medical diagnosis areas. Medical experts may use this model when accurate decision is required since their decision involves high cost or high risk. To improve our model, we should focus on developing ways to enhance both prediction accuracy and coverage ratio. In addition, methods to generate more sophisticated results should be researched in future studies. One of them may be the appropriate selection of training samples (so-called instance selection). If training samples contain noisy cases, they may distort all the parameters of our model including combination weights and thresholds. Thus, the positive effect of ‘instance selection’ on our model should be validated in future studies. Finally, the general applicability of the proposed model should be tested further by applying it to other problem domains in the future

خرید مقاله
پس از پرداخت، فوراً می توانید مقاله را دانلود فرمایید.