Poverty is a global concern that requires a significant response. With the advancement of technology, initiatives have been taken to reduce the global poverty rate. Unfortunately, due to the COVID-19 pandemic, this rate has taken another turn, greatly increasing from 8.4% in 2019 to 9.4% in 2020, with more than 33-80% of the numbers coming from India. To tackle this complex issue, many organizations have started using machine learning, which has proven to be a reliable and precise approach to curbing poverty.
By utilizing data such as communication and recharge patterns, contact networks and other records, machine learning algorithms can be used to identify poverty through three main methods. First, a machine learning model can be developed from call detail records data. Second, an asset-based index can be created to evaluate poverty levels. Lastly, a consumption metric can be used to analyze poverty situations in low and middle-income countries. It is worth noting that these algorithms utilize 797 behavioural indicators taken from CDR data.
When it comes to the accuracy of CDRs for identifying and evaluating the most deprived areas, it is compared to two other methods – assets-based and consumption-based. It has been found that CDRs achieve the highest rate of accuracy at 42%, while asset-based attain 49% and consumption-based method have 45%.
CDRs can provide a much more efficient and cheap way of collecting consumption data on a larger scale compared to traditional methods such as community-based targeting tests and proxy means tests. In addition to this, machine learning has several advantages when it comes to tackling poverty. It has the potential to revolutionize economic conditions by providing a better lifestyle for low and middle-income individuals. Nevertheless, there are drawbacks too, such as the accuracy and complexity of the solutions.
In conclusion, if the techniques of machine learning are performed well, they have the potential to eliminate or at least reduce the issue of poverty. Consequently, countries should make full use of machine learning technology and take steps to ensure that the solutions are effective and fruitful.
