| _version_ | 1866901580208406528 |
|---|---|
| author | Dr Aparna Shukla Dr. Suvendu Kanungo Dr. Vandana Bhattacharjee |
| author_facet | Dr Aparna Shukla Dr. Suvendu Kanungo Dr. Vandana Bhattacharjee |
| contents | The challenge of missing data is widespread across various domains, impacting the reliability and quality of data-driven analyses and models. Effectively addressing this challenge is imperative to uphold result integrity and facilitate accurate decision-making. The issue of missing data is a recurrent hurdle encountered in numerous research streams throughout the analysis process. Instances of substantial missing data significantly undermine the scientific reliability of causal inferences, underscoring the importance of researchers diligently addressing this concern to ensure the validity of their findings. The occurrence of missing data is influenced by diverse factors when collecting data from heterogeneous database sources, including manual data entry processes, errors in image acquisition equipment, low resolution, and other related aspects. This paper introduces an innovative method called the "Hybrid Folding Neighbour Approach" as a solution to address the challenge of missing data. This approach combines the advantages of multiple imputation techniques with a novel strategy known as neighbour folding. The paper explores various scenarios that arise due to the positioning of missing data points. By considering the location of the missing values, the concept of folding is applied to identify neighboring data points, facilitating the accurate prediction of suitable values for imputation |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_18280482 |
| institution | Zenodo |
| language | |
| publishDate | 2023 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | HFNA: A Hybrid Folding Neighbour Approach to Handle Missing Data Dr Aparna Shukla Dr. Suvendu Kanungo Dr. Vandana Bhattacharjee Data Mining; Missing Data; Imputation Method; Nearest Neighbour The challenge of missing data is widespread across various domains, impacting the reliability and quality of data-driven analyses and models. Effectively addressing this challenge is imperative to uphold result integrity and facilitate accurate decision-making. The issue of missing data is a recurrent hurdle encountered in numerous research streams throughout the analysis process. Instances of substantial missing data significantly undermine the scientific reliability of causal inferences, underscoring the importance of researchers diligently addressing this concern to ensure the validity of their findings. The occurrence of missing data is influenced by diverse factors when collecting data from heterogeneous database sources, including manual data entry processes, errors in image acquisition equipment, low resolution, and other related aspects. This paper introduces an innovative method called the "Hybrid Folding Neighbour Approach" as a solution to address the challenge of missing data. This approach combines the advantages of multiple imputation techniques with a novel strategy known as neighbour folding. The paper explores various scenarios that arise due to the positioning of missing data points. By considering the location of the missing values, the concept of folding is applied to identify neighboring data points, facilitating the accurate prediction of suitable values for imputation |
| title | HFNA: A Hybrid Folding Neighbour Approach to Handle Missing Data |
| topic | Data Mining; Missing Data; Imputation Method; Nearest Neighbour |
| url | https://doi.org/10.5281/zenodo.18280482 |