Data fraud is too perfect? too sample, GDP data more perfect!

These two days, Lynx double eleven turnover false things continue to simmer, because the curve is too perfect, it has been a lot of people questioned. Some people even use Python also a fitting operation, the fitting results and final data released Lynx is very consistent, and thus have even predicted the amount of the transaction is expected to double next year eleven, which can really surprise, the outer fly melon ah!

The figure is derived by polynomial fitting curve, it looks really perfect
Data fraud is too perfect?  too sample, GDP data more perfect!

But Lynx officials have come forward to clarify: the false data? No! must! To! Ha ha ha, it really is a matter outside the melons for the masses indistinguishable.

So when it really perfect when the data is consistent with a mathematical formula, it will be the presence of human intervention yet.
Data fraud is too perfect?  too sample, GDP data more perfect!

Since we have a sudden it difficult to distinguish between true and false Lynx data, then we can use to validate other data under way, such as GDP! This data but represents an important indicator of a country's economic strength, China's GDP that year beyond the neighbor's time, but also a fiercely proud of it!

So if we put the calendar year GDP data for some countries also fit the same way, if you have come to the perfect curve, alas, can not think, a little excited!

Let us put this fitting curve called "Lynx curve" of it, why did it body mass, the impact of the strong it!
Data fraud is too perfect?  too sample, GDP data more perfect!

Data cleansing

We are now in the hands of countries around the world over the years have a GDP data file, let's examine the first under 2018 global GDP, ranking several
Data fraud is too perfect?  too sample, GDP data more perfect!

We are in the top five for the study, and see their distribution of GDP is what

GDP historical data curve States

This top five countries, in order of priority, right on
US
Data fraud is too perfect?  too sample, GDP data more perfect!

China
Data fraud is too perfect?  too sample, GDP data more perfect!
Riben
Data fraud is too perfect?  too sample, GDP data more perfect!
Deguo
Data fraud is too perfect?  too sample, GDP data more perfect!
UK
Data fraud is too perfect?  too sample, GDP data more perfect!

This circle look down ah, seemingly in line with "Lynx curve", that is, the first two of the United States and our country, then began to fit it.

Data Fitting

About the fitting process, "Lynx curve", has written a lot of tutorials, is nothing more than polynomial regression, robust Python completely do a set, we directly see the results
Data fraud is too perfect?  too sample, GDP data more perfect!
without any problems, perfectly in line with "Lynx curve", can we say US GDP is the? Ha ha ha ha, let us look at the case again in line with the curve, in 2019 the total GDP of the United States should be

US $ 20.75 trillion

Success or failure, we wait until 2019 GDP data released then.

The following look at GDP curve situation of our country
Data fraud is too perfect?  too sample, GDP data more perfect!
do not have any accidents, more in line with the same polynomial regression curve! If we forecast 2019 GDP, it should be

US $ 14.88 trillion

Now we've got the world's highest trend curves of the two countries total GDP, we are even surprised to meet the polynomial return, is not able to explain some of the problems yet.

Radish says

We should all know, a lot of problems with the creation of mathematical formulas, are the natural scientific laws of accumulation, such as our common nonlinear regression, because there are too many phenomena in our lives, the trend data have the same law, so predecessors discovered summarizes the law of corresponding formula for unified description of these data. So when we see some of the data around is in line with normal return in the form of time, they are not much surprised.

By this time, we found the first day of the two-eleven cat transaction amount in line with polynomial return, the amount of data now contrast the United States and China's GDP, is consistent with a polynomial return, and both can be attributed to economic class data, and It can reflect the economic level of the country in certain ways. Under then we can not speculate bold, double eleven days cat transaction amount, it is China's total GDP in the domestic consumption level of some kind of reflect it. And can be seen from the growth curve in several other countries, when the large fluctuations of the national economy, not so smooth a curve, such as Japan, Germany, are after the sharp increase, in varying degrees of backward, and in the maintenance of steady economic development, the country and the United States is undoubtedly the best of the two countries do.

Wow Kaka, not a professional engaged in the economy, can not and will not say how deep, but no matter what, the data here, one thing is certain, the total amount of the country's GDP growth, people's overall standard of living will has improved, the consumption level will be a corresponding increase, among them as to whether there is some kind of special relevance, it is left to the economists to worry about it!
Data fraud is too perfect?  too sample, GDP data more perfect!

Guess you like

Origin blog.51cto.com/10487107/2450053