Acceso usuarios
Cu Nguyen Giap, Dinh Thi Ha, Dao The Son, Ngo Thi Hoai
Purpose: This study aims to apply a classification algorithm based-on fuzzy association rules (FARs) to improve the effectiveness of firms' performance prediction problem. Particularly, this study investigates potential FARs exists between inputs and outputs of firms' performance management process. These extracted FARs could be used to help firm’s managers make better dicision to improve firm’s performance.
Theoretical framework: Private enterprise development has been identified as key to Vietnam's economy that was commonly depended on state enterprise. For that, understanding and improving firms' performance and productivity is one of the most important tasks, from both macro and micro perspectives. There have been many studies on Vietnam's firm performance, but mostly relying on econometric methods that limit the understanding with structural equations. This study, instead, attempts to utilize new achievements of Artificial Intelligence (AI) for this task. Among AI techniques, fuzzy association rule is able to address the relationship between input factors and firm performance indicators. For each company, the finding FARs can be used to predict its performance and then change the business plan or react to improve weekness of organization.
Design/Methodology/Approach: The proposal model is applied on data of small and medium-sized enterprises (SMEs) of the apparel industry in Vietnam in the period 2010-2015. The sample consist of a total of 23637 observation of Vietnam firms in apparel and textile industry and contains 16 main criterias for those firms.
Finding: A recommendation system (RS) is constructed from disclosed FARs and is a key factor in a novel innovative firms' performance management process. The percentage of classified instances using the mining FARs is not quite high (about 82%), but it is not always the case. Vietnam’s apparel dataset includes rare classes of ROA, therefore applying only frequent FARs is not enough. This issue can be fixed by using both frequent and infrequent FARs. Research, practical & social implications: The proposed model has a great opportunity to use not only in the small and medium-sized enterprises (SMEs) of the apparel industry but other industrial sectors. FARs support the well-understand of firm performance to firm’s manager and help them better to react. Besides, FARs could be used to create RSs that makes alerts about risk automatically.
Originality/Value: The fact, our current study is the first to inspect the ability of FARs on SMEs of the apparel industry in Vietnam. This study provides theoritical potential knowledge and empirical evidence in the application of FARs technology in innovative firm’s management.
