Applying Clustering Technique and Association Rule to Analyze Laptop Usage Behavior of Students* Nguyen Van Chuc** [0000-0003-2618-5659] and Nguyen Ha Nhu Ngoc University of Economics, The University of Danang, Vietnam chuc.nv@due.edu.vn, ngocnhn43k08@due.udn.vn Abstract. A laptop is a typical technological product with high mobility qualities that allows everyone to learn and work from anywhere. These days, laptops are in high demand, particularly among students. There are numerous competing brands on the market with full lines, varieties, configurations, and prices ranging from inexpensive to high-end, making it difficult for customers to buy. Analyzing the behavior of students using laptops to discover trends and factors influencing their decision to buy a laptop and thus assisting them in making the best choice when choosing. It is also extremely beneficial for laptop distributors and mer- chants because it helps them to reach out to a larger number of potential custom- ers. The article focuses on applying clustering techniques and association rules in data mining to analyze the laptop usage behavior of students. Some solutions are provided based on the acquired results to assist organizations in understand- ing customer characteristics and making better business decisions. Keywords: Data Mining, Behavior Analysis, Clustering, Association Rule. 1 Introduction The Information Age is rapidly and strongly evolving, resulting in the birth of a slew of extremely modern and intelligent electronic devices. It is impossible to discuss smart devices without mentioning laptops. Recently, as a result of the global pandemic of Covid-19, there has been an increase in the use of mobile devices such as laptops for communication, distance learning, and knowledge learning based on Google’s plat- form. There are currently too many laptop lines on the market, making it difficult for users to choose the brand, function, and price that are reasonable and best suited to their personal use requirements. It’s also a common question among students. Faced with this reality, distributors and retailers must understand customer psychology, needs, and preferences in order to develop effective business policies, advertising, and marketing * Copyright © by the paper’s authors. Use permitted under Creative Commons License Attrib- ution 4.0 International (CC BY 4.0). In: N. D. Vo, O.-J. Lee, K.-H. N. Bui, H. G. Lim, H.-J. Jeon, P.-M. Nguyen, B. Q. Tuyen, J.-T. Kim, J. J. Jung, T. A. Vo (eds.): Proceedings of the 2nd International Conference on Human-centered Artificial Intelligence (Computing4Hu- man 2021), Da Nang, Viet Nam, 28-October-2021, published at http://ceur-ws.org ** Corresponding author. 68 Chuc and Ngoc strategies to increase market share and attract customers, particularly students-a poten- tial customer source for this item. 2 An overview of data clustering techniques and association rules 2.1 An introduction to clustering techniques Data clustering is the process of grouping given objects into clusters so that objects in the same cluster are as similar as possible and objects in different clusters are as differ- ent as possible. The goal of clustering is to determine the inner groupings of data. There are numerous clustering techniques available, including partition clustering, hierar- chical clustering, density-based clustering, and so on [1]. 2.2 An introduction to association rules The goal of Association Rules (AR) in data mining is to find relationships between objects in large amounts of data. The fundamental of the AR is summarized [2]. Given the transaction database T contains the set of transactions t1, t2…, tn. T = {t1, t2…, tn}. Each transaction (ti) is made up of a set of objects I (itemset). I = {i1, i2…, im}. A k-itemset is an itemset made up of k items. The purpose of AR is to discover associations (correlations) between items. These as- sociation rules take the form of X →Y, 𝑋 ∩ 𝑌 = ∅ (1) Where X: antecedents; Y: consequents Support and confidence are two crucial criteria in evaluating association rules. The for- mula for calculating the support and confidence of the association rule X → Y [2, 3]: 𝑠𝑢𝑝𝑝𝑜𝑟𝑡(𝑋 → 𝑌) = 𝑃(𝑋 ∪ 𝑌) = 𝑛 (𝑋 ∪ 𝑌)⁄𝑁 (2) 𝑐𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒(𝑋 → 𝑌) = 𝑃(𝑌|𝑋 ) = 𝑛 (𝑋 ∪ 𝑌)⁄𝑛(𝑋) (3) Where n(X): Number of transactions containing X; N: Total number of transactions. AR with support and confidence greater than or equal to the minimum support (min sup) and minimum confidence (min conf) are referred to as strong rules [2, 3]. 3 Using clustering technique and association rule to analyze student laptop usage behavior 3.1 Problem description Clustering techniques and association rules are used to analyze the laptop usage behav- ior of students. Input: information about students, laptops and factors influencing purchase. Output: provide the characteristics, behavior of using laptops across majors, predict usability, and the relationship between factors when students decide to buy a laptop. Applying Clustering Technique and Association Rule to Analyze Laptop Usage Behavior of Students 69 3.2 Implementation scenario for student laptop usage behavior analysis system Fig. 1. Implementing a student laptop behavior analysis system Step 1: Data collection and preprocessing. From March to May 2021, 1100 samples were collected through an online questionnaire survey of students from University of Economics - The University of Danang. After data preprocessing, Table 1 shows the data structure. Step 2: Deploying cluster and association rule models. The model was created using Microsoft Business Intelligence Development Studio (BIDS) data mining tool. The dis- covered knowledge is very intuitive, easy to understand, and simple to apply [4, 5]. Step 3: The knowledge discovered from the Cluster and Association rule models used to analyze laptop usage behaviour of students. (Fig.1) Table 1. Data description No. Attribute name Data types Value domain Meaning 1 ID Interval 1 - 1100 Individual survey order 2 GioiTinh Nominal Male, Female Gender of the students 3 QueQuan Nominal Da Nang, Hue… The student’s hometown 4 NamHoc Nominal 1st, 2nd year Year students are studying 5 Khoa Nominal Banking… Faculty 6 Nganh Nominal Accounting… Major 7 NgheNghiepGD Nominal Farmers… Parents’ occupations 8 ChiTieuHangThang Nominal 1.5 million VND Monthly budget for a student 9 ThuongHieu Nominal Asus, Dell… Brand of laptop 10 ThoiGianMua Nominal Under 6 months When purchasing a laptop 11 Gia Nominal 15 million VND Laptop pricing 12 MucDichSuDung Nominal Studying, … Purpose of buying a laptop 13 MucDoHaiLong Interval 1→5 Level of satisfaction 14 ThongTinMua Nominal Websites… Sources of information 15 NoiMua Nominal FPT Shop ... Laptop stores 16 YeuToThuongHieu Nominal Very important... Factor of Brand 70 Chuc and Ngoc 17 YeuToCauHinh Nominal Important… Factor of Configuration 18 YeuToTocDoXuLy Nominal Normal... Factor of Processing Speed 19 YeuToGiaCa Nominal Unimportant... Factor of Price 20 YeuToKieuDang Nominal Normal… Factor of Style 21 YeuToUyTin Nominal Very important... Factor of Retailer Reputation 22 YeuToBaoMat Nominal Very important... Factor of Confidential Mode 23 YeuToBaoHanh Nominal Very important... Factor of Warranty 24 YeuToKhuyenMai Nominal Very important... Factor of Promotion Fig. 2. The results of data clustering According to the model’s results (Fig.2), Table 2 shows six clusters: Table 2. Cluster characteristics Cluster Size Cluster characteristics Female in 3rd year, parents are - Configuration, Processing speed: farmers, the cost between 10 Normal; Brand, Warranty, Promo- 259 Cluster 1 and 15 million VND. Seeking tion, Price, Retailer reputation, (23.5%) information from relatives and Style: Important; Security: Unim- friends, expert → Dell, Asus. portant. Female in 4th year, parents are - Processing speed, Style: Normal; farmers, and the cost between Brand, Configuration, Warranty, Se- 210 Cluster 2 10 and 15 million, according to curity, Promotion, Retailer reputa- (19.1%) information obtained from rel- tion: Important; Price: Very im- atives and friends → Dell. portant. Female in 2nd, 3rd years, par- ents are government employ- - Promotion: Normal; Price, Config- 196 ees, self-employed. It cost over uration, Processing speed, Warranty, Cluster 3 (17.8%) 25 million VND to seek infor- Security: Important; Brand, Retailer mation from family and sales- reputation, Style: Very important. people → Apple. 151 Male in 2nd and 3rd years, par- - Brand, Warranty, Security, Promo- Cluster 4 (13.7%) ents are farmers, and the cost tion, Retailer reputation: Normal; Applying Clustering Technique and Association Rule to Analyze Laptop Usage Behavior of Students 71 from 10 to 15 million VND, Price, Configuration, Processing seeking purchasing information speed: Important; Style: Unim- salespeople → Dell and Asus. portant. Male in 3rd year, parents are farmers and self-employed. - Security: Normal; Promotion, The cost between 15 and 20 Brand, Warranty, Price, Retailer rep- 132 Cluster 5 million VND. They seek pur- utation: Important; Style: Unim- (12.0%) chasing information from portant; Configuration, Processing salespeople and expert guid- speed: Very important. ance → Dell and HP. Female in 2nd year, parents are - Processing speed: Normal; Brand, farmers with laptop prices un- Warranty, Price, Retailer reputation: 152 der 10 million VND, referring Cluster 6 Important; Security, Style, Configu- (13.8%) to buying information from ration: Unimportant; Promotion: family, friends, and salespeo- Very important. ple → Asus and Acer. Fig. 3. The results of association rule model. Here are some association rules (Fig.3): R1: With a price of 25 million VND, second-year Foreign Trade students in Da Nang are interested in brand design and security issues when purchasing a laptop that primar- ily uses the Apple brand, with 100 percent confidence. R2: Students of Commerce and Management Information Systems (MIS) from Quang Nam, Ha Tinh, are interested in warranty, configuration, and brand factors when purchasing a laptop with a price range of 10 to 15 million VND. Dell is the most com- monly used brand, with 75 percent confidence. R3: With a price range of 10 to 15 million VND, third-year Accounting students in Dak Lak, who are concerned with price, promotion, and warranty when purchasing a laptop, primarily use the Asus brand, with 63 percent confidence. R4: With a price of less than 10 million VND, Accounting and Commerce students in Quang Nam and Hue, are interested in the price, promotion, and warranty factors when purchasing laptops that primarily use the Asus and Acer brands, with 60 percent confidence. 72 Chuc and Ngoc R5: With a price range of 15 to 20 million VND, a male student majoring in MIS whose parents are freelancer, is interested in configuration factors, processing speed, and warranty when purchasing HP laptops, with 53 percent confidence. 4 Proposing data-driven marketing and CRM solutions 4.1 Describing product feature The clustering results show that Dell and Asus are two brands that students frequently purchase, with prices ranging from 10 to 15 million VND. To minimize shortages, dis- tributors should concentrate on these two important brands. The findings of the association rule highlight some aspects of various majors: About the brand The majority of Foreign Trade students utilize Apple products; Students of MIS like the HP brands, which cost 15-20 million VND; Students of Accounting and Business choose Asus and Acer brands with prices under 10 million VND. About the factors Foreign Trade students are interested in design and security factors; MIS students are interested in configuration and processing speed factors; Accounting, Business and Commerce students are interested in price and promotion factors. 4.2 Sources of information for purchasing a laptop Improve image promotion, coverage and word-of-mouth marketing: Use the credibility of celebrities and loyal customer groups. Enhance product quality and the company’s reputation; develop after-sales customer care plans to encourage customers. - Organize the hiring and training of professional staff who are familiar with laptops and have a professional service style. Training and coaching staff to always smile at customers in all situations. - Create a separate customer service department to handle customer feedback. All re- quests for purchasing, selling and delivering services are handled in a timely manner. Companies can also set up separate phone lines to handle customer inquiries. 4.3 Programs for advertising and marketing Companies must improve their communication on social networking sites, as well as websites that combine programs, fairs, and exhibitions held at schools. Updating infor- mation on e-commerce channels to disseminate knowledge and general information about product features. According to the clustering results, students are subjects with a limited budget who place a high value on price and promotion. As a result, distribution and retail companies should: Applying Clustering Technique and Association Rule to Analyze Laptop Usage Behavior of Students 73 - Implement a marketing strategy suitable to students of different majors according to their behaviours and characteristics. - Offer optimal payment terms, such as implementing an installment policy, purchas- ing first and paying later with a 0% interest rate. - Put product discounts, promotions, and giveaways into action. Promotions must be diverse and of high quality. - There are policies in place to allow customers to return goods and receive refunds in certain circumstances, such as when laptop malfunctions. 5 Conclusions and future work The article studied the theory of clustering techniques and association rules for applying these techniques to build data mining models to analyze the laptop usage behavior of students. The results of the clustering technique analysis have clarified the outstanding features of groups of students with similar characteristics; the association rules discov- ered from the data help to understand the relationship and influence of factors influenc- ing students’ choice to buy laptops. The knowledge extracted from the models assists laptop distributors and retailers in understanding the trends and characteristics of stu- dents, allowing them to develop effective business strategies. More data will be gath- ered in the coming months from a variety of sources, including not only students from University of Economics-The University of Danang, but also extensive research with students from all over Danang City, making the data more complete, in order to improve the model and increase the efficiency of analysis and prediction. References 1. Chuc, N.V. and Giang, D.T.: Applying clustering technique and association rule to exploit data of customers using hotel services. Journal of science and technology (JST-UD), 12(97), 109-112 (2015). http://tapchikhcn.udn.vn/volume/135/Ung_dung_ky_thuat_phan_cum_va_luat_ket_hop_k hai_pha_du_lieu_khach_hang_su_dung_dich_vu_khach_san-8501/8501 2. Tan, P.N, Steinbach M., Karpatne, A., Kumar V.: Introduction to data mining, 2 nd Edition, Pearson (2018). 3. Witten, L.H, Frank, E., Hall, M.A., Pal, C.J.: Data mining: Practical machine learning tools and techniques, 4th Edition, Morgan Kaufmann (2016). 4. SQL Server Analysis Services (SSAS). https://docs.microsoft.com/en-us/analysis-ser- vices/data-mining/data-mining-ssas?view=asallproducts-allversions. 5. Data Mining and Business Intelligence solution. http://bis.net.vn/forums/36.aspx