Chapter 797 Beat
After all, the scale of his companies and the industries they span have surpassed Delong.
And compared to the Tang brothers, he is only twenty-six years old.
It’s the age when ‘you can’t do anything well if you don’t talk well’.
When others see him for the first time, they can't help but become suspicious.
Can such a young person really integrate the huge Delong Group?
Is such a young person really in charge of Delong Assets, Hongmeng, Hanhua, and Master Kong?
Once he goes astray like Old Tang, the impact of the collapse of the Xu Group will be ten times or a hundred times that of Delong.
The first assistant is not worried.
However, this inherent impression based on age is a habit formed over thousands of years, and Xu Liang alone cannot change it.
He doesn't want to change either.
It would be great if the enemy looked down upon him because of this.
While Xu Liang continued to read the speech materials he had prepared, a staff member came over.
"Mr. Xu, the time is up, please come with me."
Xu Liang nodded.
After briefly sorting out the information, he strode out the door.
Under the leadership of the staff, we came to a hall.
After the other party made a gesture to invite him in, Xu Liang stopped, took a deep breath, and strode in.
‘Crash’!
Loud applause sounded.
This warm applause also dispelled a trace of nervousness in Xu Liang's heart.
Standing on the rostrum, he glanced down.
Sixty or seventy people were seated on the rows of sofas. The head of state and the chief minister were sitting at the front, surrounded by members of the elders.
Xu Liang saw the figure of his old father-in-law behind them.
From now on, except for a very few, the rest are basically unknown.
He looked down at the laptop in front of him.
The PPT inside has been moved to the first page.
Open the speech you brought in and put it on the table.
After calming down his emotions, Hong said.
“I am honored to be invited by the cabinet to give this speech. I am a little nervous to face so many big names for the first time. If there is something wrong with what I say next, I hope all seniors will consider me. For the sake of a yellow-haired boy, please forgive me.”
"Xiao Xu, feel free to speak boldly.
Don't have any baggage in your heart, just treat us as your employees." The head of state smiled.
"With your words, I feel relieved."
After saying a polite word, Xu Liang stopped wasting time.
“The title of today’s speech is: The impact of big data, cloud computing and artificial intelligence on the future!
First, let’s be clear, what is data?
In many people’s impressions, numbers are data, or they must be composed of numbers.
In fact, data is much bigger than numbers.
Any content on the Internet, such as text, pictures and videos, is data.
All files in the hospital, including medical images, are also data;
Various design drawings in companies and factories are also data;
The texts and illustrations on unearthed cultural relics, and even their sizes and materials, are also data.
Even our human activities themselves can be regarded as a special kind of data.
Data in various fields around the world continue to expand outwards, and gradually form another feature, that is, a lot of data begins to intersect.
Data in various dimensions gradually evolved from points and lines into a network.
In other words, the correlation between data has been greatly enhanced. In this context, big data has emerged. "
After a pause, Xu Liang adjusted the PPT.
"So how to use data and big data?
It can be roughly divided into the following processes.
Get data → analyze data → build model → predict the unknown.
Let's take a simple example.
Now we want to know the age distribution of a cinema’s audience for marketing purposes.
Suppose we divide the audience into four groups: under 15 years old, 16~25 years old, 26~40 years old and 41 years old and above.
To understand the proportion of each group of people, a simple way is to go to the entrance of the cinema and ask the ages of those watching the movie.
For example, through our survey, we learned that there are about 343 people under 15 years old, 459 people between 16 and 25 years old, 386 people between 26 and 40 years old, and 490 people at 41 years old and above.
Based on this data, we can roughly draw the following conclusions:
Audiences aged 15 and under account for about 20%, and audiences aged 16 to 25 account for more than a quarter, but less than 30%;
< br>Slightly less than a quarter of the audience is between the ages of 26 and 40, and the audience aged 41 and above is the largest, accounting for about 30%.
But if we only sample 10 people on weekend nights, we will find out.
There are three spectators aged 15 and under, five spectators aged 16~25, and two spectators aged 26~40.
We obviously cannot conclude that 80% of the audience is under the age of 25, while middle-aged people aged 41 and above never come to the cinema.
But I think everyone also admits that when the statistical samples are insufficient, the results obtained will deviate greatly from the actual results.
So, the more you want to get accurate statistical results, the greater the amount of statistical data you need.
In the above example, the total number of samples counted is 1678 people.
But if we must say ‘audiences aged 41 and above are 29.2%’, or ‘audiences aged 15 and under must exceed 20%’.
If it is so certain, everyone may challenge this conclusion. Because statistics are random and have errors.
Only data from thousands of people cannot produce such an accurate conclusion.
In addition to requiring sufficient data volume, statistics also requires that the sampled data must be representative.
Sometimes, if the amount of data is not large enough, it will be accurate if the same level is used.
A very simple example, a love film and a war film have different audiences.
So if we only collect the moviegoers in the month when the romance movie is released, it will not be universally representative.
So how to avoid this situation and obtain accurate conclusions?
The 19th century Russian mathematician Chebyshev gave his conclusion on this problem, namely Chebyshev's inequality.
P(|X-E(X)|≥ε)≤Var(X)/ε^2.
The meaning of this formula is that when the number of samples is large enough, the error between a random variable and its mathematical expectation value can be arbitrarily small.
Apply Chebyshev's inequality to our problem of understanding the age distribution of movie theater audiences.
The random variable is: the observed proportion of viewers of each age group.
The mathematical expectation value is: the proportion of different age groups among all moviegoers under real circumstances.
When we bring in the sample data, we can roughly draw the following conclusions.
Audiences under 15 years old accounted for 20%, 16 to 25 years old accounted for 27%, 26 to 40 years old accounted for 24%, and over 40 years old accounted for 29%, with an error of less than 5%.
But if we want to improve the accuracy of the four age groups of viewers to one digit after the decimal point, then we need roughly 10 times the data, that is, about 20,000 samples.
If we magnify this problem.
We want to know the age distribution of a movie’s moviegoers around the world, and it must be specific to a more detailed age group.
For example, 18~20 years old, 21~24 years old, etc.
Or a more specific region.
China, Japan, South Korea, etc.
In a larger and more detailed range, in order to obtain more accurate results, the amount of data we need will increase hundreds of times.
When we get super data.
It is already difficult for ordinary computers to complete calculations.
And even if it can be completed, it will take a lot of time.
Time is money, and in business, this is obviously unacceptable.
Therefore.
In order to get the results as quickly as possible, we need one or several supercomputers to calculate.
But using supercomputers is very expensive.
Companies that want to know the age of movie theater audiences are obviously unwilling to spend such a high price on this issue.
So what to do? ”
Xu Liang operated the computer.
Three huge characters in regular script were displayed on the projection screen behind him.
Cloud computing.
"Cloud computing, 'cloud' is the Internet, and 'computing' is the literal meaning.
The current cloud computing is a kind of distributed computing, which refers to the calculation and processing of huge data through the network "cloud" The program is decomposed into countless small programs.
Then, these small programs are processed and analyzed through a system composed of multiple servers, and the results are returned to the user.
The entire calculation process is just one step. It takes a few seconds.
In other words, cloud computing solves a problem that would have taken several days or even ten days to solve using a supercomputer.
It has become accurate data that can be obtained in just a few seconds and at a cost of tens of thousands or at most hundreds of thousands of dollars.
Significantly reduce expenses, improve efficiency, and get more accurate results.
Some people may think that statistics on the age distribution of movie audiences are of little value.
But what if it becomes a catering company to calculate the age distribution of the audience for beverage products?
As long as they have accurate data, catering companies can develop more targeted advertisements and services for people of different age groups, thereby increasing their sales.
This point has been applied by Master Kong in actual operations.
According to Pangu Company’s big data survey, Master Kong found that the largest audience for its ‘Jianlibao’ sports drink product is teenagers between the ages of 15 and 25.
Among them, 41% are men and 59% are women.
Then the data was obtained through offline sampling surveys, online questionnaires and other channels, and data-driven methods were used to calculate the celebrities that this group of people are interested in, the types of TV series they like, etc.
Synthesizing these intersecting big data, a detailed advertising plan and publicity channels were formulated.
In just one quarter, Master Kong's sales increased by 22% and net profit increased by 14.8%.
The same method can be applied to all consumer goods fields such as automobiles, restaurants, and entertainment.
There is no doubt that this will form a huge business change.
The original extensive advertising and publicity methods will become more detailed and targeted.
The original products with uniform taste will be developed to be more in line with local characteristics based on the tastes and consumption habits of consumer groups in various provinces across the country.
Consumers will become the real subjects.
It can be said with certainty that all consumer goods companies that reject big data in the future will basically not survive for long. ”
(End of this chapter)