A method of credit evaluation modeling based on block-wise missing data

Lan, Qiujun; Jiang, Shan

doi:10.1007/s10489-021-02225-5

Cited by 5 publications

(9 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This has the advantage that less parameters need to be estimated for the iSFS, which should lead to more stable models. Lan and Jiang (2021) reformulate the iMSF method as a multi-task learning problem. Liu et al (2017) again divide the observations into subsets based on the missingness pattern in a similar way to iMSF, where, however, the available data are exploited better than for iMSF.…”

Section: Methods That Deal With the Missingness Patternmentioning

confidence: 99%

See 1 more Smart Citation

Prediction approaches for partly missing multi-omics covariate data: A literature review and an empirical comparison study

Hornung¹,

Ludwigs²,

Hagenberg³

et al. 2023

Preprint

View full text Add to dashboard Cite

As the availability of omics data has increased in the last few years, more multi-omics data have been generated, that is, high-dimensional molecular data consisting of several types such as genomic, transcriptomic, or proteomic data, all obtained from the same patients. Such data lend themselves to being used as covariates in automatic outcome prediction because each omics type may contribute unique information, possibly improving predictions compared to using only one omics data type. Frequently, however, in the training data and the data to which automatic prediction rules should be applied, the test data, the different omics data types are not available for all patients. We refer to this type of data as block-wise missing multi-omics data. First, we provide a literature review on existing prediction methods applicable to such data. Subsequently, using a collection of 13 publicly available multi-omics data sets, we compare the predictive performances of several of these approaches for different block-wise missingness patterns. Finally, we discuss the results of this empirical comparison study and draw some tentative conclusions.

show abstract

Section: Methods That Deal With the Missingness Patternmentioning

confidence: 99%

“…The methods by Chen and Zhang (2020), Ingalhalikar et al (2012), and Lan and Jiang (2021) were not designed for the multi-omics case. However, they can still be applied to such data.…”

Section: Methods That Deal With the Missingness Patternmentioning

confidence: 99%

Prediction approaches for partly missing multi-omics covariate data: A literature review and an empirical comparison study

Hornung¹,

Ludwigs²,

Hagenberg³

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Multi-source random forests (Ludwigs, 2020), iMSF (Yuan et al, 2012), and MMPFS (Q. Lan & Jiang, 2021) are the only methods that do not need the prediction rule to be retrained when new test data sets with missing values are obtained. As described above, with multi-source random forests the trees are pruned to use only covariates that are available in the training data.…”

Section: Methods That Deal With the Missingness Patternmentioning

confidence: 99%

“…Q. Lan and Jiang (2021) reformulate the iMSF method as a multi-task learning problem in their method MMPFS. Liu et al (2017) again divide the observations into subsets based on the missingness pattern in a similar way to iMSF, where, however, the available data are exploited better than for iMSF.…”

Section: Methods That Deal With the Missingness Patternmentioning

confidence: 99%

Prediction approaches for partly missing multi‐omics covariate data: A literature review and an empirical comparison study

Hornung

Ludwigs

Hagenberg

et al. 2023

WIREs Computational Stats

View full text Add to dashboard Cite

As the availability of omics data has increased in the last few years, more multi‐omics data have been generated, that is, high‐dimensional molecular data consisting of several types such as genomic, transcriptomic, or proteomic data, all obtained from the same patients. Such data lend themselves to being used as covariates in automatic outcome prediction because each omics type may contribute unique information, possibly improving predictions compared to using only one omics data type. Frequently, however, in the training data and the data to which automatic prediction rules should be applied, the test data, the different omics data types are not available for all patients. We refer to this type of data as block‐wise missing multi‐omics data. First, we provide a literature review on existing prediction methods applicable to such data. Subsequently, using a collection of 13 publicly available multi‐omics data sets, we compare the predictive performances of several of these approaches for different block‐wise missingness patterns. Finally, we discuss the results of this empirical comparison study and draw some tentative conclusions.This article is categorized under: Applications of Computational Statistics > Genomics/Proteomics/Genetics Applications of Computational Statistics > Health and Medical Data/Informatics Statistical and Graphical Methods of Data Analysis > Analysis of High Dimensional Data

show abstract

“…The purpose of model selection is to determine a model with fk that best fits the data with 'n' sub-models as shown in Eq. (7) for the k th model, where, 1 ≤ i ≤ n. The model averaging technique combines more than one feasible model that are built with weight updation. Let w 1 , w 2 , ...., w n denote the weights assigned to the multiple sub-models of the model averaging estimator [Sun et.…”

Section: Model Estimations With Weighted Probabilitymentioning

confidence: 99%

A Machine Learning Algorithm for Multi-Source Heterogeneous Data With Block-Wise Missing Information

Ghorpade-Aher¹,

Sonkamble²

2022

INDJCSE

View full text Add to dashboard Cite

The heterogeneous data coming from varied sources is integrated to induce learning as it may have complementary information. The growing research for multi-source heterogeneous data analysis has extended the traditional methods of single source homogeneous data. One of the major challenges is processing the heterogeneous data while integrating the information from different sources with various missing data patterns. The potential features act as the useful predictors, but rejecting the observations with missing data values will lose the information which may create biased conclusions affecting the model strength for uniform decisions. A general imputation technique should be preferred for its computational simplicity and capability to induce a diminutive bias in the dataset. The proposed 'Heterogeneous Data and Weight Algorithm' constructs the robust model averaging techniques and feasible estimations with advanced Machine Learning techniques that implements the weighted probability approach to enhance the performance of the model with better predictive power.

show abstract

A method of credit evaluation modeling based on block-wise missing data

Cited by 5 publications

References 48 publications

Prediction approaches for partly missing multi-omics covariate data: A literature review and an empirical comparison study

Prediction approaches for partly missing multi-omics covariate data: A literature review and an empirical comparison study

Prediction approaches for partly missing multi‐omics covariate data: A literature review and an empirical comparison study

A Machine Learning Algorithm for Multi-Source Heterogeneous Data With Block-Wise Missing Information

Contact Info

Product

Resources

About