fedirt_2PL_median_data {FedIRT}

Federated 2PL model

Description

This function is only used to test the accuracy and processing time of this algorithm. It inputs a list of responding matrices and return the federated 2PL parameters. Note: This function can only calculate one combined dataset. To use federated 2PL in distributed datasets, please use fedirt_2PL().

Usage

fedirt_2PL_median_data(inputdata)

Arguments

inputdata

A List of all responding matrix.

Details

Input is a List of responding matrices from each school, every responding matrix is one site's data. It uses Federated median instead of FedAvg and FedSGD. The results are not same as traditional 2PL method, but is robust if there are outliers.

Value

A list with the estimated global discrimination a, global difficulty b, person's abilities ability, sites' abilities site, and log-likelihood value loglik.

Examples

inputdata = list(as.matrix(example_data_2PL))
fedresult = fedirt_2PL_median_data(inputdata)

inputdata = list(as.matrix(example_data_2PL_1), as.matrix(example_data_2PL_2))
fedresult = fedirt_2PL_median_data(inputdata)