Data dimension reduction and decentralization
According to reports, blockchain has the advantages of decentralization, distrust and tamper proof
compared with the Internet, the blockchain network conceived by Cosmos not only realizes information interaction, but also realizes the transmission of asset value. Through the IBC cross chain protocol, the token can be transferred across the chain between the blockchains based on Cosmos' tendermint core. For the public chains based on POW consensus mechanism like Ethereum, pegged zone can be used to bridge
Article source: bit110 network
method
has multiple perspectives on the role of data dimensionality rection. Wu Enda said in his video that dimensionality rection is used for data compression to rece noise and prevent slow running and small memory; When it is reced to 2 or 3 dimensions, it can be visualized for data analysis; Don't use dimension rection to prevent over fitting. It's easy to remove important features related to tags. But why data need to be compressed, in addition to occupying memory, is there any other reason - "dimension disaster" problem: the higher the dimension, the more sparse the distribution of your data on each feature dimension, which is basically disastrous for machine learning algorithms. The final result may be that each sample has its own characteristics, which can not form a unified feature to distinguish positive cases from negative cases. There is another case, when the feature is more than the sample size, some classification algorithms (SVM) are invalid, which is related to the principle of classification algorithm<
data dimension rection method:
& 65532<
linear dimensionality rection method:
principal component analysis (PCA) and discriminant analysis (LDA)
understanding of PCA:
1. PCA can be understood as the projection of high-dimensional data to low-dimensional data, and make the projection error minimum. It's an unsupervised method
2. It can also be understood as coordinate rotation and translation (corresponding to coordinate transformation and decentralization), so that the n-dimensional space can be analyzed in n-1 dimension, and the characteristics of small variance (small variance, small uncertainty, small amount of information)
3. Derivation of PCA
4. Connection between PCA and SVD
(Understanding PCA from the perspective of matrix decomposition)
5. Application of PCA dimension rection
6 Disadvantages of PCA:
(1) PCA is a linear dimensionality rection method, sometimes the nonlinear relationship between data is very important, when we use PCA, we will get very poor results. Next, we introce PCA of kernel method
(2) principal component analysis is more effective only when the sample points obey Gaussian distribution
(3) cost sensitive PCA (cspca) can be used to rece the dimension of imbalanced data.
(4) the size of feature roots determines how much information we are interested in. In other words, small feature roots often represent noise, but in fact, the projection to smaller feature roots may also include the data we are interested in
(5) the directions of eigenvectors are orthogonal, which makes PCA vulnerable to outlier
(6) it is difficult to explain the results. For example, in the establishment of linear regression model (linear regression model) analysis of dependent variables
2. Many characteristics of data greatly affect the effect of neural network and other classification models. For example, the distribution of data value is not on the same scale, and the local temperature value and the local monthly wage are obviously not on the same order of magnitude. At this time, the data need to be normalized, and the data of these two characteristics should be normalized to 0 to 1, so that their impact on the model has the same scale
3. Of course, there are many data preprocessing, such as centralization, denoising, dimensionality rection, smoothing, transformation and so on, which have their own purposes. In short, they are all for the final classifier effect service. Since the original data may contain a lot of noise, it is necessary to remove the noise.
"blockchain reform" and "blockchain transformation" started in 2018. On August 5, 2018, the chain reform action plan [1] was first released at the second plenary session of the block chain special committee of China Communications Instry Association, and since then, "chain reform" has become a hot word in the circle. The more common areas of chain reform include finance, right protection, games, government certificates, etc. Since then, the term "chain reform" has been widely used in news reports. Chain reform has also been widely used in more instries, covering finance, social networking, e-commerce, games, media, retail, energy, agriculture, ecation, medical and beauty, medical and other instries
in August, the opinions of the CPC Central Committee and the State Council on supporting Shenzhen to build a leading demonstration zone of socialism with Chinese characteristics [2] proposed to support the development of innovative applications such as digital currency research and mobile payment in Shenzhen. Experts believe that as it takes time for the launch, trial operation and improvement of digital currency, the pilot program should be launched "sooner rather than later". Shenzhen has the foundation to carry out pilot projects. The introction of digital currency will greatly promote the development of cross-border trade, especially cross-border trade financing. This major positive is undoubtedly another benign stimulus to the chain reform
on October 24, the 18th collective learning meeting on the current situation and trend of blockchain technology development emphasized that the integrated application of blockchain technology plays an important role in new technological innovation and instrial change. It is necessary to take blockchain as an important breakthrough for independent innovation of core technology, clarify the main direction, increase investment, strive to conquer a number of key core technologies, and accelerate the innovation and development of blockchain technology and instry[ 3] This collective learning is undoubtedly a great benefit for the whole blockchain instry. It can promote the faster, more stable and healthier development of the instry from the national level, and is a boost to the confidence of the practitioners in the blockchain instry[ 4]
after the speech of the general secretary, relevant learning activities were carried out all over the country. In Cheng, representatives of relevant enterprises such as block chain professional committee of China Communications Instry Association, Sichuan strategic emerging instry promotion association, chain Expo (Cheng) science and technology attended the study meeting [5]; Similar learning activities have been held in Beijing, Hangzhou and other places. It can be seen that the chain reform has been paid attention to at the national level
Apple 7 plus is a al 12 megapixel camera with natural virtual effect. You can see the preview of background virtual in the photo interface. Brightness increased by 25%, better color
the mobile phone has 12 million pixels. Among the 3362 smart phones, the camera pixels are more than 47% of the mobile phones
iPhone 7 plus is an upgraded version of Apple 7, which was released at Apple's autumn new proct launch on September 8, 2016 Beijing time
the iPhone 7 plus has a 5.5-inch screen size and al 12 megapixel cameras. The virtual effect is natural, the brightness is increased by 25%, and the color is better. Speaker upgrade, using up and down stereo speakers. The 3.5mm headphone connector was canceled and a new apple earpods was introced
an appointment was opened on September 9, 2016, and it was officially put on sale on September 16, 2016, starting in China. The iPhone 7 plus starts at 6388 yuan. There are silver, gold, rose gold, black, bright black and red. There are 32GB, 128GB and 256gb optional light in the storage. The outside lens is broken, which generally has little impact on the imaging. If there is no crack in the photo, it will not affect the pixel. The principle is different. If you change all the photos together, it will not be too expensive. A few hundred yuan. The key is to see if the photos have any impact. The iPhone 6S uses a 1200W pixel camera, It can't be very bad, because I use 6S. I think it's possible for you to buy a fake or break the lens. I suggest you take it to Apple's after-sales point to have a look, or send me the serial number. I'll help you check whether it's genuine (enter * # 06 # to get the serial number at the dial-up) thank you
in the second half of 2019, major mobile phone manufacturers are making efforts one after another, Launch its own annual flagship procts to attract consumers' attention. In 5g is not yet fully popular, folding screen technology is not fully mature today, camera configuration is undoubtedly the main driving point of innovation. A single camera head has long been out of date. Now it is the world of three and four cameras, and even some manufacturers have launched 100 megapixel mobile phones
all major brands boast about their procts. Which is better? Recently, DxOMark updated the rating list of mobile cameras to tell us the answer. Briefly, this is a well-known authoritative camera evaluation organization abroad. They evaluate the total score by testing camera exposure, color, texture, noise, zoom, night shooting ability and other factors. The mobile phones that can be on the list are shooting
first, instrial restructuring: the instrial logic behind the nationwide entrepreneurship tide, the Internet is evolving from the first generation to the second generation, from a trillion level market space to dozens of trillion level market space:
the first generation Internet is an era dominated by portals, games, search and social networking, which originated from the technological revolution, supply creates demand, The core business model is cash flow, which creates a trillion scale game & advertising market online. Up to now, it has begun to enter a normal but not sexy state. The first generation Internet belongs to the geek world and follows the winner take all rule
when Jingdong was once the biggest traffic company in the world, and LETV, A-share company, attacked the traffic leader, people found that the simple traffic logic failed
we are entering the second generation Internet! The Internet has become an information energy and has begun to reshape the relationship between supply and demand in the real society. The business model has also evolved from simple flow cash to two directions: rising to cloud and big data and sinking to o2o. The extended instrial chain makes the Internet from the geek world to the whole people
what enterprises have opportunities? We believe that the Rising Cloud and big data are the game of the giant ecosystem, and bat has long understood it; The sinking o2o will bring a lot of opportunities for the rise of leading companies in new fields. We do not want to list all trades and the Internet simply, which is confusing, impatient and confused. Instead, we hope to find out certain rules and directions, and truly understand this "Internet plus" revolution before we can really enjoy the whole capital feast. For the first time, we propose to find the enterprises that have the best chance to win in the second generation Internet according to the attributes of two different instries:
1) instries with high traditional concentration and channel monopoly premium will usher in Internet disruptive innovation. The leaders of such traditional instries enjoy channel monopoly and brand premium, and the companies in the instry have no incentive to break the existing balance“ Success will become the epitaph of success ", and the Internet will eliminate the channel / brand cost of the original instrial chain in the way of dimension rection, so as to provide services to the demand side more efficiently. In this case, the Internet will subvert the instry, so-called "disruptive innovation" and "decentralization". Such as Xiaomi, Jingdong, LETV; If the enterprises in the original instry participate in the reform, they will face the embarrassing situation of "left and right mutual understanding", and it is difficult to succeed; Internet companies with knowledge of the instry have no historical burden, but have more advantages. After a round of harvesting of traditional manufacturers, Internet companies that have a firm foothold in traditional instries will form a higher monopoly, and the accumulation of data in traditional instries will even give these new leaders the opportunity to develop to the cloud and big data mode of the second generation Internet. Finally, a giant with more than 100 billion US dollars was born
2 the original leader of "heart plug instry" will get more opportunities. There are many pain points in the original service, opaque information, lack of credit system, or policy restrictions, which lead to the immature development and low concentration of the whole instry. This kind of instry is called "Xinsai instry", and the typical ones are decoration, finance, automobile after service, ecation, medical care, etc. Its primary demand is to use the Internet to improve the pain points of the original instry. Because the cost of learning the Internet for the original instry olfactory person is lower than that for the pure Internet person, the winning enterprise is often the leader of the original instry. Such enterprises have been working in traditional fields for many years, fully understand the pain points of the instry, have strong offline promotion, strong back-end capabilities, and are assisted by Internet tools, and will get greater opportunities. If we say that the objective reasons in the original traditional field make the enterprise unable to grow, now the excellent management team + advantageous capital card + Internet will make the enterprise energy exponentially enlarged. There are many such companies in the current A shares, and they are facing the opportunity of Jedi reversal
after the reconstruction of the Internet, China will usher in the next dividend after the demographic dividend big data dividend, leading the world Internet experience upgrading trend. It can be optimistically predicted that if the center of manufacturing instry is in Germany and the center of innovation is in the United States, then the ultimate center of next generation Internet Data & service is in China
two, capital reconstruction: benign bubbles stimulate innovation, and foster the fundamental driving force of economic transformation. This will be the home court of RMB capital feast. At present, in the transitional economy of China, the first round of real instrial capital needs to be exported. Emerge in an endless stream of two most of the market, the Internet is the most promising instry direction. So the reaction in the market is more intense. So the bubble of speech and fear is emerging. We believe that this dramatic wealth effect is not a bad thing. Instead, it will stimulate innovation, encourage the whole people to start businesses, and eventually cultivate the mode of economic transformation. Different from the first generation of Internet promoted by US dollar capital, RMB capital's personal perception and in-depth understanding of instrial development will give more momentum to the second generation of Internet. At the same time, the relatively clear business model of the second generation Internet will not completely repeat China's US internet bubble. The mentality of instrial capital and the personnel structure of the capital market decide that the secondary market will take the lead in enjoying the capital feast< Third, valuation Restructuring: bringing three-stage investment opportunities. Shopping spree is a long-term and solid investment in the Internet plus. We believe that although the technology will fluctuate in the short run, the long-term Internet market is firmly committed to the "Internet plus" bull market. The current Internet investment will go through three stages:
1. At present, we are in the stage of "shopping for chips". The Internet supply is unable to meet the hot pursuit demand of capital.
2) "other people laugh at me for being too crazy, but I laugh at them for not seeing through". Since there was no real first generation Internet company before the secondary market of A-share, we are still learning from the first round of Internet era. When the U.S. - led high-tech companies entered the valuation system of PV, UV and dau, the real leaders who had already run out of the track were drowned by revelry and panic, In the absence of a consensus on the safety margin, a short break after the carnival will provide investors with an excellent opportunity to get involved
3) "you have been identified as a male and female for a long time, and detailed tracking is the most important". The former innovative enterprises have become "traditional" and entered the stage of stable development, and the valuation has reached the consensus expectation
we judge that before and after the implementation of the registration system, A-shares may graally enter the second stage. With the qualitative VC valuation method, companies will be selected from the three dimensions of market space, enterprise card and management team, and they will enjoy more than 10 times of growth space if they firmly buy and hold for a long time.
Why can't it be explained from the perspective of traffic distribution? Because you can see that bat, these giants, have mastered the channels and made a lot of money by distributing traffic and providing information. But I don't see why these giants can control the channel and get the right to distribute traffic. Bat's layout on the Internet covers all aspects of our lives - shopping, payment, catering, entertainment, transportation, communication, social networking, etc.. However, most of them can be summarized in one word -
intermediary
intermediary has already penetrated into all aspects of our life for a long time. Almost, as long as there is a certain demand, as long as the information between supply and demand is not equal, or even a little inconvenient, in the era of Internet, there will be intermediary. Once the intermediary was expensive and charged according to the service, so there were only some common intermediaries in the instry. However, in the Internet era, especially for Internet giants, the income from data and channels can completely make up for the intermediary costs, which is why several Internet giants are now saying that they must be ecological, and they always have layout...
the reason why internet giants are able to distribute traffic and provide information is that they have mastered the channels, leading to traffic, In other words, information can only be distributed through them. In other words, the distribution of traffic becomes a centralized process, not because the information or traffic itself is centralized, but because bat centralizes the distribution of information through layout and providing intermediary services, and then the center is in their hands
take any example, such as takeout. No meituan, Internet, hungry? Haven't we had takeout before? No, we can also know the take out calls of several small shops nearby. This is a decentralized system. Ordering such things does not require the participation of a third party. But why is this field now completely centralized? The reason is that these companies provide intermediaries and a centralized platform. At the same time, they provide some things that could not be provided by the previous decentralized system, such as convenient information for query and use, (relatively) real evaluation and scoring system, and convenient payment system (this is another centralized system)
as a result, when bat arranges these companies, bat will control this channel. Since then, the ordering information will almost certainly be distributed through their platform, which seems to give people the illusion of "bat makes money by distributing ordering information". However, don't forget that the reason why they are able to master this channel and the right to distribute the information is that they provide this intermediary system, destroy the original decentralized system with poor information exchange and opacity, and make the instry become the current situation that all information needs to be distributed through them
the reason why you think "changing the underlying system is just changing the traffic center for distributing information" is because you think "this is a centralized system, so there must be a center". But I forget that these fields originally had no center, because someone first provided this centralized system and destroyed the original decentralized system, so this field had a center
this is also the reason why this article and people are optimistic about blockchain -
all the things that these intermediary systems can do, blockchain can also do (or have the potential to do), and the functions of these Internet Intermediary platforms can be achieved through algorithms without bat or any Internet giants
maybe now, the resistance of technology, capital and other aspects makes the blockchain not allow to destroy these intermediaries as the centralized platform destroys the traditional decentralized market, but the blockchain provides this possibility, which is why many people say that the blockchain is revolutionary, or that it can change the Internet
as for the reality, can we kill bat? I can only say that the current blockchain has almost no ability to kill anything. Let's not say that a centralized system like meituan has been formed. If we leave the blockchain in an area that bat has not yet touched or disdains to touch, the blockchain must first kill the original decentralized system and say something else
at present, as far as I know, blockchain hasn't got a single shot
the first generation Internet is the era dominated by portals, games, search and social networking. Due to the technological revolution, supply creates demand, and the core business model is flow realization, Online to create a trillion scale space game & advertising market. Up to now, it has begun to enter a normal but not sexy state. The first generation Internet belongs to the geek world and follows the winner take all rule
when Jingdong was once the biggest traffic company in the world, and LETV, A-share company, attacked the traffic leader, people found that the simple traffic logic failed
we are entering the second generation Internet! The Internet has become an information energy and has begun to reshape the relationship between supply and demand in the real society. The business model has also evolved from simple flow cash to two directions: rising to cloud and big data and sinking to o2o. The extended instrial chain makes the Internet from the geek world to the whole people
what enterprises have opportunities? We believe that the Rising Cloud and big data are the game of the giant ecosystem, and bat has long understood it; The sinking o2o will bring a lot of opportunities for the rise of leading companies in new fields. We do not want to list all trades and the Internet simply, which is confusing, impatient and confused. Instead, we hope to find out certain rules and directions, and truly understand this "Internet plus" revolution before we can really enjoy the whole capital feast. For the first time, we propose to find the most promising enterprises in the second generation Internet according to the attributes of two different instries:
1) instries with high traditional concentration and channel monopoly premium will usher in Internet disruptive innovation. The leaders of such traditional instries enjoy channel monopoly and brand premium, and the companies in the instry have no incentive to break the existing balance“ Success will become the epitaph of success ", and the Internet will eliminate the channel / brand cost of the original instrial chain in the way of dimension rection, so as to provide services to the demand side more efficiently. In this case, the Internet will subvert the instry, so-called "disruptive innovation" and "decentralization". Such as Xiaomi, Jingdong, LETV; If the enterprises in the original instry participate in the reform, they will face the embarrassing situation of "left and right mutual understanding", and it is difficult to succeed; Internet companies with knowledge of the instry have no historical burden, but have more advantages. After a round of harvesting of traditional manufacturers, Internet companies that have a firm foothold in traditional instries will form a higher monopoly, and the accumulation of data in traditional instries will even give these new leaders the opportunity to develop to the cloud and big data mode of the second generation Internet. Finally, a giant with more than 100 billion US dollars was born
2 the original leader of "heart plug instry" will get more opportunities. There are many pain points in the original service, opaque information, lack of credit system, or policy restrictions, which lead to the immature development and low concentration of the whole instry. This kind of instry is called "Xinsai instry", and the typical ones are decoration, finance, automobile after service, ecation, medical care, etc. Its primary demand is to use the Internet to improve the pain points of the original instry. Because the cost of learning the Internet for the original instry olfactory person is lower than that for the pure Internet person, the winning enterprise is often the leader of the original instry. Such enterprises have been working in traditional fields for many years, fully understand the pain points of the instry, have strong offline promotion, strong back-end capabilities, and are assisted by Internet tools, and will get greater opportunities. If we say that the objective reasons in the original traditional field make the enterprise unable to grow, now the excellent management team + advantageous capital card + Internet will make the enterprise energy exponentially enlarged. There are many such companies in the current A shares, and they are facing the opportunity of Jedi reversal
after the reconstruction of the Internet, China will usher in the next dividend after the demographic dividend big data dividend, leading the world Internet experience upgrading trend. It can be optimistically predicted that if the center of manufacturing instry is in Germany and the center of innovation is in the United States, then the ultimate center of next generation Internet Data & service is in China
two, capital reconstruction: benign bubbles stimulate innovation, and foster the fundamental driving force of economic transformation. This will be the home court of RMB capital feast. At present, in the transitional economy of China, the first round of real instrial capital needs to be exported. Emerge in an endless stream of two most of the market, the Internet is the most promising instry direction. So the reaction in the market is more intense. So the bubble of speech and fear is emerging. We believe that this dramatic wealth effect is not a bad thing. Instead, it will stimulate innovation, encourage the whole people to start businesses, and eventually cultivate the mode of economic transformation. Different from the first generation of Internet promoted by US dollar capital, RMB capital's personal perception and in-depth understanding of instrial development will give more momentum to the second generation of Internet. At the same time, the relatively clear business model of the second generation Internet will not completely repeat China's US internet bubble. The mentality of instrial capital and the personnel structure of the capital market decide that the secondary market will take the lead in enjoying the capital feast
Third, valuation Restructuring: bringing three-stage investment opportunities. Shopping spree is a long-term and solid investment in the Internet plus. We believe that although the technology will fluctuate in the short run, the Internet will be able to go through three stages in the long term. 1, we are currently in the stage of "shopping for chips". The scarce supply of Internet targets can not meet the hot pursuit of capital, the popularity of the target and the carnival of the plate. 2) "People laugh at me for being too crazy, but I laugh at them for not seeing through it." because there was no real first generation Internet company before the secondary market of A-share, we were still learning from the first round of Internet era. When the US led high-tech companies entered the valuation system of PV, UV and dau, the real leaders who had already run out of the track were drowned by revelry and panic, In the absence of a consensus on the safety margin, a short break after the carnival will provide investors with an excellent opportunity to get involved. 3) "It has long been recognized that you are male and female, and detailed tracking is the most important." once innovative enterprises have become "traditional" and entered a stage of stable development, and the valuation has reached consensus expectations
we judge that before and after the implementation of the registration system, A-shares may graally enter the second stage. By means of qualitative VC valuation, the company will select tomorrow's leading companies from three dimensions of market space, enterprise card and management team, firmly buy and hold for a long time, and will enjoy more than 10 times of growth space.