Poverty Classification Using Machine Learning: The Case of Jordan

Abstract

The scope of this paper is focused on the multidimensional poverty problem in Jordan. Household expenditure and income surveys provide data that are used for identifying and measuring the poverty status of Jordanian households. However, carrying out such surveys is hard, time consuming, and expensive. Machine learning could revolutionize this process. The contribution of this work is the proposal of an original machine learning approach to assess and monitor the poverty status of Jordanian households. This approach takes into account all the household expenditure and income surveys that took place since the early beginning of the new millennium. This approach is accurate, inexpensive, and makes poverty identification cheaper and much closer to real-time. Data preprocessing and handling imbalanced data are major parts of this work. Various machine learning classification models are applied. The LightGBM algorithm has achieved the best performance with 81% F1-Score. The final machine learning classification model could transform efforts to track and target poverty across the country. This work demonstrates how powerful and versatile machine learning can be, and hence, it promotes for adoption across many domains in both the private sector and government.

Sustainable FinTech Innovation Orientation: A Moderated Model

“…On the other hand, other respondents live in the northern and southern regions, representing 77 and 51 individuals, respectively (see Table 1). Jordan was chosen as the study's focus because it is at the forefront of technological innovation, providing tangible support to institutions and technology businesses alike and attempting to establish a regulatory framework that will assist in technology drive and build Jordan's economy [38]. Jordan is one of the Middle East's developing countries.…”

Deploying digitalisation and artificial intelligence in sustainable development research

“…The proposed solution was tested on the real data – the satellite images presenting the area affected by the Tohoku tsunami in November 2010 Sustainable extreme phenomena damages management Sublime and Kalinicheva ( 2019 ) 6 RecycleNet: intelligent waste sorting using deep neural networks The RecycleNet is a carefully optimised deep convolutional neural network architecture to classify selected recyclable object classes: glass, paper, cardboard, plastic, metal, and trash AI-supported waste management Bircanoğlu et al ( 2018 ) 7 Poverty classification using machine learning: the case of Jordan Proposal of a machine learning approach to assess and monitor the poverty status of Jordanian households Better tracking and targeting poverty across the country. The work demonstrates how powerful and versatile machine learning can be, enabling its adoption across many private and government domains Alsharkawi et al ( 2021 ) 8 Sustainability assessment and modelling based on supervised machine learning techniques: the case for food consumption Presents a method for evaluating and modelling sustainability impacts of food consumption in the United States through the assessment of categories by (1) using high sector resolution input–output of the economy and (2) proposing an integrated sustainability modelling framework based on supervised machine-learning techniques The supervised machine-learning techniques allows to develop sustainability modelling and assessment method that deals with multiple decision-making units (food consumption categories) and sustainability indicators Abdella et al ( 2020 ) 9 Combining satellite imagery and machine learning to predict poverty Machine learning techniques and scalable method to predict poverty by estimating consumption expenditure and asset wealth from high-resolution satellite imagery and survey data; The proposed method and machine learning techniques can foster research and policy Jean et al ( 2016 ) 10 Artificial intelligence-enhanced decision support for informing global sustainable development: a human-centric AI-thinking approach Democratisation of AI via a user-friendly human-centric probabilistic reasoning approach and the application of AI-based predictive modelling techniques on Environmental Performance Index data, revealing tensions between (1) environmental health; and (2) ecosystem vitality...…”

“…Despite the attempts that try to tackle sustainability problems by adopting a more systemic and interdisciplinary perspective, several other studies address the AI and digitalisation related methods to advance specific SD challenges such as urban water management (Goralski & Tan, 2020 ; Xiang et al, 2021 ), disaster detection (Al Qundus et al, 2020 ; Alizadeh & Nikoo, 2018 ; Sublime & Kalinicheva, 2019 ), recycling and waste sorting (Bircanoğlu et al, 2018 ), poverty assessment and monitoring (Alsharkawi et al, 2021 ; Jean et al, 2016 ), food consumption and sustainability (Abdella et al, 2020 ), sustainable cities/urban ecosystems (Goddard et al, 2021 ; Ilieva & McPhearson, 2018 ; Majumdar et al, 2021 ; Vinuesa et al, 2020 ), threat to biodiversity (Jensen et al, 2020 ), health-related issues (Pirouz et al, 2020 ), smart sustainable agriculture (Alreshidi, 2019 ; Boev et al, 2020 ), climate action (Balogun et al, 2020 ; Fuso Nerini, Slob, et al, 2019 ; Fuso Nerini, Sovacool, et al, 2019 ; Huntingford et al, 2019 ), environmental evaluation (Liu et al, 2021 ), and enterprise management for SD (Kościelniak et al, 2019 ).…”

“…In some ways, the expert opinions obtained through the survey reflect several key points identified in the relevant literature. First, of all systems associated with AI, machine learning tends to be one of the most promising in the context of SD research (Abdella et al, 2020 ; Alsharkawi et al, 2021 ; Huntingford et al, 2019 ; Jean et al, 2016 ; Jensen et al, 2020 ; Majumdar et al, 2021 ). Second, substantial agreement exists to suggest that AI and digitalisation have clear applications to the study of smart grids (Aly, 2020 ; Kotsiopoulos et al, 2021 ; Liang et al, 2020 ; Qadir et al, 2021 ; Shi et al, 2020 ; Talaat et al, 2020 ) and smart and low-carbon cities (Fuso Nerini, Slob, et al, 2019 ; Fuso Nerini, Sovacool, et al, 2019 ; Goddard et al, 2021 ; Ilieva & McPhearson, 2018 ; Majumdar et al, 2021 ).…”

Diagnosis of Diabetes Mellitus Using Gradient Boosting Machine (LightGBM)

“…Nowadays, for classification and diagnosis problems, LightGBM outperforms other state-of-the-art methods, cf. [33][34][35][36][37][38][39][40]. In these related works, LightGBM is not only selected for its effective prediction performance, but also for its shorter computational time and optimized data handling technique.…”