Answers are grounded in this paper and cite their sources.
This paper aims to examine the data requirements and predictive modeling methodologies for estimating broadband take-up ratio (TUR).The author applies the CRISP-DM approach to define data requirements based on the results of the analysis of the company's internal as well as external factors.Using Kitchenham approach, the author applies the findings of the literature review on machine learning predictive model to the second study area.The CRISP-DM strategy has been demonstrated through the creation of data needs based on business imperatives, data understanding activities that include defining data analysis requirements, identification of attributes for internal and external variables as well as determining resource sources available in a database information system.In addition, as indicated in the system data sources, seven internal factors and nineteen external factors have been identified within the Customer Management System, Survey System, Relations or Micro Clusters.Consequently, neural networks are superior to SVMs and XG Boost when analyzing the data to predict how much uptake there will be.The Neural Network model performed best with an AUC of 79-89, an accuracy level of 75.23-99, an F Measure of 80.97-95.35,an F1 Score of 94.5-98.5, and an R 2 score of 88.SVM, with an AUC of 98.8, accuracy of 92. is rated as a prediction model in second place.In contrast, XGBoost had an AUC of 77-99.1,accuracy of 77-98.9, and F1 score of 98.3 as the 3 rd .In the case of TUR classification use, XG Boost and SVM were also becoming increasingly common.In terms of AUC, accuracy, F measure and F1 score, XGBoost is more successful than SVM.The AUC of the XG Boost is 99.1, an accuracy range of 77-98.6 and an F1 value of 98.3.Although SVM's AUC is almost the same at 98.8, its F1 score is lower than XG Boost.In conclusion, the research contributes to the identification of relevant data sources and provides an initial framework for the selection of the TUR model methodology.To establish the optimal approaches, additional studies should focus on robust testing and validation of models.
No takes yet. Share an insight, caveat, or question.
Haris et al. (2024) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: