The Impact of Feature Selection: A Data‐mining Application in Direct Marketing

Tan, Ding-Wen; Yeoh, William; Boo, Yee Ling; Liew, Soung-Yue

doi:10.1002/isaf.1335

Cited by 6 publications

(4 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In other words, DM has the potential to transform traditional SMEs to organizations with a competitive advantage. For instance, suppose an SME aims to mine its customer data, the potential benefit would entail creating cross-selling avenues at a higher margin, improving its customer retention and satisfaction rates, identifying the most profitable customer group and last but not least, enhancing the SME's marketing and sales strategy (Tan, Yeoh, Boo, & Liew, 2013). In the mining of inventory data, on the other hand, SMEs can gain an advantage by forecasting the inventory that will help reduce the total value of stock held.…”

Section: What Dm Can Mean For Smes?mentioning

confidence: 99%

Big data analytics—A review of data‐mining models for small and medium enterprises in the transportation sector

Selamat

Prakoonwit

Sahandi

et al. 2018

WIREs Data Min & Knowl

View full text Add to dashboard Cite

The need for small and medium enterprises (SMEs) to adopt data analytics has reached a critical point, given the surge of data implied by the advancement of technology. Despite data mining (DM) being widely used in the transportation sector, it is staggering to note that there are minimal research case studies being done on the application of DM by SMEs, specifically in the transportation sector. From the extensive review conducted, the three most common DM models used by large enterprises in the transportation sector are identified, namely “Knowledge Discovery in Database,” “Sample, Explore, Modify, Model and Assess” (SEMMA), and “CRoss Industry Standard Process for Data Mining” (CRISP‐DM). The same finding was revealed in the SMEs' context across the various industries. It was also uncovered that among the three models, CRISP‐DM had been widely applied commercially. However, despite CRISP‐DM being the de facto DM model in practice, a study carried out to assess the strengths and weakness of the models reveals that they have several limitations with respect to SMEs. This paper concludes that there is a critical need for a novel model to be developed in order to cater to the SMEs' prerequisite, especially so in the transportation sector context. This article is categorized under: Application Areas > Business and Industry Application Areas > Industry Specific Applications

show abstract

Section: What Dm Can Mean For Smes?mentioning

confidence: 99%

Big data analytics—A review of data‐mining models for small and medium enterprises in the transportation sector

Selamat

Prakoonwit

Sahandi

et al. 2018

WIREs Data Min & Knowl

View full text Add to dashboard Cite

show abstract

“…Thus, several studies were published related to the application of feature selection to CT [17]. A recent study authored by Tan et al [18] shows the impact of feature selection in direct marketing. Their work analysed three filter methods for feature selection (correlation-based feature selection, subset consistency and symmetrical uncertainty), concluding that symmetrical uncertainty resulted in better models, outperforming both the two remaining methods studied and a model without any feature selection procedure.…”

Section: Introductionmentioning

confidence: 99%

“…While there are several studies published on feature selection using MI, and a few using SA, none performed a direct comparison on both methods to assess the pros and cons on using each. Furthermore, even though a handful of recent studies were found comparing feature selection methods through practical applications [18], none considered SA. The main contributions of this study are as follows:…”

Section: Introductionmentioning

confidence: 99%

Mutual information and sensitivity analysis for feature selection in customer targeting: A comparative study

Barraza

Moro

Ferreyra³

et al. 2018

Journal of Information Science

View full text Add to dashboard Cite

Feature selection is a highly relevant task in any data-driven knowledge discovery project. The present research focus on analysing the advantages and disadvantages of using mutual information (MI) and data-based sensitivity analysis (DSA) for feature selection in classification problems, by applying both to a bank telemarketing case. A logistic regression model is built on the tuned set of features identified by each of the two techniques as the most influencing set of features on the success of a telemarketing contact, in a total of 13 features for MI and 9 for DSA. The latter performs better for lower values of false positives while the former is slightly better for a higher false positive ratio. Thus, MI becomes a better choice if the intention is reducing slightly the cost of contacts without risking losing a high number of successes. On the other side, DSA achieved good prediction results with less features.

show abstract

“…Through an extensive conducted it was uncovered that the hindrance factors of data analytics adoption are in the area of data management, knowledge management and data management [9]. Data Mining (DM) is cru-cial for SMEs, especially so, in the private transport domain as the organisation can inherently extract and process its dataset to uncover new insights to facilitate better decision for the business [10,11]. In an extensive study of DM application by the large transportation enterprise and SMEs context at large, it was uncovered that the CRISP-DM methodology is the most popular applied DM model by the industry [9].…”

Section: Introductionmentioning

confidence: 99%

Enhanced CRISP-DM model for SME coach operator

2018

J. adv. tec. eng. res.

View full text Add to dashboard Cite

The surging call for SMEs in the private transportation domain to apply data analytics in its business is evidential. However, to date, there is a lack of speci ic data mining methodology or framework customized to meet the demand of the private transportation domain. Considered as a de-facto data mining methodology by the industry to date, the Cross-Industry Standard Process for Data Mining (CRISP-DM) consists of a six-phase process that is extendable in the framework -from generic to specialized tasks. The traditional CRISP-DM methodology exempli ies the DM application area, issues identi ications, technical, tools and technique requirements. Through this study, an Enhanced CRISP-DM for SME Coach Operator (ECSMCO) methodology was developed. The extended methodology aims to curb the existing application limitation identi ied in the small and medium-sized enterprises (SMEs) by proposing an enhancement to the existing CRISP-DM activities. To evaluate the novel ECSMCO's acceptability, three UK SME coach operators were identi ied to apply and evaluate the ECSMCO methodology. The outcome of this paper will ascertain the users' acceptability outcome of the applied ECMSCO methodology.

show abstract

The Impact of Feature Selection: A Data‐mining Application in Direct Marketing

Cited by 6 publications

References 16 publications

Big data analytics—A review of data‐mining models for small and medium enterprises in the transportation sector

Big data analytics—A review of data‐mining models for small and medium enterprises in the transportation sector

Mutual information and sensitivity analysis for feature selection in customer targeting: A comparative study

Enhanced CRISP-DM model for SME coach operator

Contact Info

Product

Resources

About