In machine learning problems, high dimensional data, especially in terms of many features, is increasingly these days [1]. Many researchers focus on the experiment to solve these problems. Besides, to extract important features from these high dimensional of variables and data. The statistical techniques were used to minimize noise and redundant data. Nevertheless, we do not use all the features to train a model. We may improve our model with the features correlated and non-redundant, so feature selection plays an important role.

Purpose Zakat during the COVID-19 outbreak has played a vital role and has been significantly discussed in the virtual environment. Such information about zakat in the virtual world creates unstructured data, which contains important information and knowledge. This paper aims to discover knowledge related to zakat administration during the pandemic from the information in a virtual environment. Furthermore, the discussion is contextualised to the socio-economic debates. Design/methodology/approach This is a qualitative study operated via text mining to discover knowledge of zakat administration during the COVID-19 pandemic. The National Board of Zakat Republic of Indonesia (BAZNAS RI) is selected for a single case study. This paper samples BAZNAS RI’s situation report on COVID-19 from its virtual website. The data consists of 40 digital pages containing 19,812 characters, 3,004 words and 3,003 white spaces. The text mining analytical steps are performed via RStudio. The following R packages, networkD3, igraph, ggraph and ggplot2 are used to run the Latent Dirichlet Allocation (LDA) for topic modelling. Findings The machine learning analysis via RStudio results in the 16 topics associated with the 3 primary topics (i.e. Education, Sadaqah and Health Services). The topic modelling discovers knowledge about BAZNAS RI’s assistance for COVID-19 relief, which may help the readers understand zakat administration in times of the pandemic from BAZNAS RI’s virtual website. This finding may draw the theory of socio-economic zakat, which explains that zakat as a religious obligation plays a critical role in shaping a Muslim community's social and economic processes, notably during the unprecedented times of COVID-19. Research limitations/implications This study uses data from a single zakat institution. Thus, the generalisation of the finding is limited to the sampled institution. Practical implications This research is both theoretically and practically important for academics and industry professionals. This paper contributes to the novelty in performing text mining via R in gaining knowledge about the recent zakat administration from a virtual website. The finding of this study (i.e. the topic modelling) is practically essential for zakat stakeholders to understand the contribution of zakat in managing the COVID-19 impacts. Social implications This work derives a theory of “socio-economic zakat” that explains the importance of a zakat institution in activating zakat for managing socio-economic issues during the pandemic. Thus, paying zakat to an authorised institution may actualise more maslahah (public interest) compared to paying it directly to the asnaf (zakat beneficiaries) without any measurement Originality/value This study is among the pioneers in gaining knowledge from Indonesia’s zakat management during the COVID-19 outbreak via text mining. The authors’ way of analysing data from the virtual website using RStudio can advance Islamic economics literature.

The H-likelihood method proposed by Lee and Nelder (1996) is extensively used in a wide range of data. In terms of direction, repetitive measured data within classification can be examined employing hierarchical generalized linear models (HGLMs). Whether we are concerned in multiple endpoints which are correlated, instead Multivariate Double Hierarchical Generalized Linear Models (DHGLM) can be taken into consideration. This paper addresses the implementation of this principle to vector selection and support machines. Based on the analysis with the fish morphology class Sardinella lemuru (Bali sardinella) and setting the best epsilon 0.7 cost 4 parameter reaching best performance: 0.2327401. Predictive value of fish sex was calculated 0.997319 and Region under the curve: 0.8967. At the same time, we extend the large-scale case studies for stress testing of the SVM method by using three datasets from UCI machine learning repository including the Bank marketing dataset, the car evaluation database and Human Activity Recognition Using Smartphones dataset. In a nutshell by employing SVM-DHGLM increased the accuracy, precision, recall, for feature selection and classification. Long story short, the H-likelihood provides an excellent and usable structure for statistical inference of the unobservable general deterministic model, while preserving the advantages of the original probability structure for fixed parameters. We presume that more new groups of models will be created and that the H-likelihood will be commonly used for their inferences and the application in big data and machine learning.