Information Entropy-Based Data Pricing

In recent years,the upsurge of Mobile Internet and Internet of Things(Io T)technology has brought human society into an era of big data,spawning a brand-new industry-data industry.The scale of data transactions is growing at an alarming rate.Data has been traded the way like products in general terms,but confronted a question about how to be reasonably priced.The pricing theory and strategy of general merchandise and information merchandise could not guide the pricing of data goods well.On the other hand,the current research on the pricing of data goods is still in its infancy,with only some pricing mechanisms based on data exchange platform scenarios.It is noted that there is no pricing mechanism that considers information amount contained in the data itself as a pricing factor.Besides,is it justified that data seller sells out the ownership of the data? Because data is a special class of product with high sunk cost and low marginal cost.To solve the above problems,a data pricing research based on information entropy has been conducted in this thesis.Firstly,a new data pricing metric,Data Information Entropy,is proposed to measure the information amount contained in a dataset.The measurement method of data information entropy is given.In addition,this thesis also defines a pricing function based on data information entropy and discusses some good economic properties of the pricing function.In order to make the proposed pricing function applicable,this thesis further explores the possibility of combining data entropy with auction mechanism.Firstly,the data online auction based on data information entropy is discussed,and a second-value online auction model based on data information entropy is proposed.Finally,a new scenario of data online transaction is defined,where data platforms sell the right of use of their data products through auctions.The main contributions of this thesis are summarized as follows:1.The traditional pricing theories ignore the impact of data information amount on its price,and information amount is obviously an important metric of data products value;2.Proposed a method to measure the information amount of dataset based on information entropy,and verified the rationality of data information entropy through a large number of experiments;3.Proposed a pricing function based on data information entropy,which has good economic properties,such as no arbitrage helping to avoid the fraud in the trading platform.4.Explored the possibility of combining the online auction with the data information entropy.A second-value online auction model based on data information entropy has been proposed.The expected revenue of the seller and expected transaction price of the model have been analyzed;5.Defined a new online data auction scenario.Besides,several guidances have been proposed for online auction mechanism design.Data trading is a huge potential and lucrative industry,and data pricing is the basis for trading.In order to seize the opportunities of the data trading industry and improve the theory and method of modern information economics,this thesis has researched about information entropybased data pricing,in an expectation to better price data based on its information amount.We hope this thesis can start the research of data pricing based on data information entropy,and bring new ideas to data traders,thus enhancing the betterment of data economy.