It so appears to be the case that satellite surveillance of our planet allows scientists all the better to keep track of poverty. The regions associated with poverty got identified.
The latest method of predicting poverty is through satellite imagery. Also machine learning comes in handy in this regard. This technique could make it a cinch for philanthropic groups to gather valuable data.
Don't Miss: iPhone 8: Everything You Need to Know
They can then pinpoint the regions where they ought to be concentrating all their efforts at poverty alleviation.
Governments too can start making worthier policies and plans regarding the erasure of poverty. One sign of an area’s affluence is how well lit up it is at night.
The more the external lights, the more well off are the denizens living in that particular loci. This methodology is employed in areas about which we do not have precise data.
However, there is a catch to all this. Night light rates are a rough guesstimate and they are not as accurate as they were once supposed to be. The differences between the rich and the poor are much more complex than just street lights and lamp posts.
Scientists at Stanford University entered the following data into a computer: night light pics, daylight images and real survey information. This was so that an algorithm could be developed which would allow prediction regarding the wealth or poverty status of a region.
The results were published in the journal Science.
Poverty in the backward regions of the world is generally tough to measure. Economic markers such as household income and fixed assets come in handy as indicators of one’s pecuniary status.
The only issue is that much of this data is unavailable. It is too expensive to collect all the tons and tons of information in order to pinpoint the places of destitution in the global village.
Instead of night lights, it is daytime imagery that may come to the rescue in this conundrum. Machine learning is the key component here. Five African countries had their daytime and night time images taken via satellite.
The daytime features which would translate into night time lights were determined. Areas which were naturally dark at night such as lakes were deleted from the poverty statistics.
A ridge regression model was also used. The rest of the steps taken were quite complex and involved heuristics and computer models.
Ultimately, after a lot of analysis, transfer learning was employed to calculate the relative poverty or wealth of a region. As for its accuracy, it worked just fine as a representative model.
Buy Now: Sony PlaysStation VR In Stock Here