Displaying all 3 publications

Abstract:
Sort:
  1. Fulltext Examining the relationship between big data analytics capabilities and organizational ambidexterity in the Malaysian banking sector
    Aziz NA, Long F
    Front Big Data, 2023;6:1036174.
    PMID: 37007906 DOI: 10.3389/fdata.2023.1036174
    Drawing on previous literature on dynamic capability view (DCV), we examine the effects of data analytics capabilities (BDAC) on organizational ambidexterity and the paradoxical tensions between exploration and exploitation in the Malaysian banking sector. Although banks are often considered as mature commercial organizations, they are not free of issues concerning technological advancement and organizational changes for long-term competitiveness. Through statistical analysis by using data from 162 bank managers in Malaysia, it is confirmed that BDAC positively influences the two contradictory aspects of organizational ambidexterity (i.e., explorative dynamic capabilities and exploitative dynamic capabilities), and explorative dynamic capabilities also mediate the positive relationship between BDAC and exploitative marketing capabilities. The findings provide meaningful insights to researchers and bank managers on how to obtain sustainable competitive advances in the current digital era.
  2. Fulltext Survival Analysis of COPD Patients in a 13-Year Nationwide Cohort Study of the Brazilian National Health System
    Gargano LP, Zuppo IF, do Nascimento MMG, Augusto VM, Godman B, Costa JO, et al.
    Front Big Data, 2021;4:788268.
    PMID: 35198972 DOI: 10.3389/fdata.2021.788268
    BACKGROUND: Chronic obstructive pulmonary disease (COPD) has an appreciable socioeconomical impact in low- and middle-income countries, but most epidemiological data originate from high-income countries. For this reason, it is especially important to understand survival and factors associated with survival in COPD patients in these countries.

    OBJECTIVE: To assess survival of COPD patients in Brazil, to identify risk factors associated with overall survival, including treatment options funded by the Brazilian National Health System (SUS).

    METHODOLOGY: We built a retrospective cohort study of patients dispensed COPD treatment in SUS, from 2003 to 2015 using a National Database created from the record linkage of administrative databases. We further matched patients 1:1 based on sex, age and year of entry to assess the effect of the medicines on patient survival. We used the Kaplan-Meier method to estimate overall survival of patients, and Cox's model of proportional risks to assess risk factors.

    RESULT: Thirty seven thousand and nine hundred and thirty eight patients were included. Patient's survival rates at 1 and 10 years were 97.6% (CI 95% 97.4-97.8) and 83.1% (CI 95% 81.9-84.3), respectively. The multivariate analysis showed that male patients, over 65 years old and underweight had an increased risk of death. Therapeutic regimens containing a bronchodilator in a free dose along with a fixed-dose combination of corticosteroid and bronchodilator seem to be a protective factor when compared to other regimens.

    CONCLUSION: Our findings contribute to the knowledge of COPD patients' profile, survival rate and related risk factors, providing new evidence that supports the debate about pharmacological therapy and healthcare of these patients.

  3. Design of a data processing method for the farmland environmental monitoring based on improved Spark components
    Tang R, Aridas NK, Talip MSA
    Front Big Data, 2023;6:1282352.
    PMID: 38053722 DOI: 10.3389/fdata.2023.1282352
    With the popularization of big data technology, agricultural data processing systems have become more intelligent. In this study, a data processing method for farmland environmental monitoring based on improved Spark components is designed. It introduces the FAST-Join (Join critical filtering sampling partition optimization) algorithm in the Spark component for equivalence association query optimization to improve the operating efficiency of the Spark component and cluster. The experimental results show that the amount of data written and read in Shuffle by Spark optimized by the FAST-join algorithm only accounts for 0.958 and 1.384% of the original data volume on average, and the calculation speed is 202.11% faster than the original. The average data processing time and occupied memory size of the Spark cluster are reduced by 128.22 and 76.75% compared with the originals. It also compared the cluster performance of the FAST-join and Equi-join algorithms. The Spark cluster optimized by the FAST-join algorithm reduced the processing time and occupied memory size by an average of 68.74 and 37.80% compared with the Equi-join algorithm, which shows that the FAST-join algorithm can effectively improve the efficiency of inter-data table querying and cluster computing.
Related Terms
    Try searching for something
Filters
Contact Us

Please provide feedback to Administrator (afdal@afpm.org.my)

External Links