Petrônio Cândido de Lima e Silva, Carlos Alberto Severiano, Marcos Antônio Alves, Rodrigo Silva · 6 authors
In this paper we introduce a Non-Stationary Fuzzy Time Series (NSFTS) method with time varying parameters adapted from the distribution of the data. In this approach, we employ Non-Stationary Fuzzy Sets, in which perturbation functions are used to adapt the membership function parameters in the knowledge base in response to statistical changes in the time series. The proposed method is capable of dynamically adapting its fuzzy sets to reflect the changes in the stochastic process based on the residual errors, without the need to retraining the model. This method can handle non-stationary and heteroskedastic data as well as scenarios with concept-drift. The proposed approach allows the model to be trained only once and remain useful long after while keeping reasonable accuracy. The flexibility of the method by means of computational experiments was tested with eight synthetic non-stationary time series data with several kinds of concept drifts, four real market indices (Dow Jones, NASDAQ, SP500 and TAIEX), three real FOREX pairs (EUR-USD, EUR-GBP, GBP-USD), and two real cryptocoins exchange rates (Bitcoin-USD and Ethereum-USD). As competitor models the Time Variant fuzzy time series and the Incremental Ensemble were used, these are two of the major approaches for handling non-stationary data sets. Non-parametric tests are employed to check the significance of the results. The proposed method shows resilience to concept drift, by adapting parameters of the model, while preserving the symbolic structure of the knowledge base.
A blockchain, such as Bitcoin, is an append-only, secure, transparent, distributed ledger. A fair blockchain is expected to have healthy metrics; high honest mining power, low processing latency, i.e., low wait times for transactions and stable price of consumption, i.e., the minimum transaction fee required to have a transaction processed. As Bitcoin matures, the influx of transactions increases and the block rewards become insignificant. We show that under these conditions, it becomes hard to maintain the health of the blockchain. In Bitcoin, under these mature operating conditions (MOC), the miners would find it challenging to cover their mining costs as there would be no more revenue from merely mining a block. It may cause miners not to continue mining, threatening the blockchain's security. Further, as we show in this paper using simulations, the cost of acting in favor of the health of the blockchain, under MOC, is very high in Bitcoin, causing all miners to process transactions greedily. It leads to stranded transactions, i.e., transactions offering low transaction fees, experiencing unreasonably high processing latency. To make matters worse, a compounding effect of these stranded transactions is the rising price of consumption. Such phenomena not only induce unfairness as experienced by the miners and the users but also deteriorate the health of the blockchain. We propose BitcoinF transaction processing protocol, a simple, yet highly effective modification to the existing Bitcoin protocol to fix these issues of unfairness. BitcoinF resolves these issues of unfairness while preserving the ability of the users to express urgency and have their transactions prioritized.
Sofia Santiago Marinho, José Silveira Filho, Leonardo O. Moreira, Javam C. Machado
Relational Databases (RDBs) have been widely used for decades. However, new persistence technologies are emerging, such as Blockchain, which is disruptive and has relevant properties, such as immutability and no third parties. Therefore, applications that use RDB can benefit from these properties by migrating part of their data to Blockchains. This article presents the MOON, a hybrid approach to manage data in RDB and Blockchain, which receives SQL queries. A case study was performed with a real health dataset using three scenarios. The conclusion is that the MOON responds to requests correctly and provides RDB and Blockchain features. Moreover, its response time was intermediate between RDB and Blockchains.
Currently, there are hundreds of Bitcoin exchanges on the market, so choosing a reliable exchange is a critical issue for users. We know that the amount of Bitcoin holdings is an essential indicator for evaluating an exchange, but people have very few ways to access this information. Besides, many reports indicate that the trading volumes of most Bitcoin exchanges do not match their real situations, and the fake volume has become an unspoken rule of the whole industry. It causes the public to doubt the actual amount of Bitcoin owned by each exchange. To solve the problem of information asymmetry between users and exchanges, we propose a method for tagging Bitcoin addresses of exchanges. Through vertical, forward, and backward address mining, the method can utilize only one or several addresses of an exchange to find out all its addresses and distinguish different address types: deposit wallet, hot wallet, and cold wallet. Then the balance and transfers of the exchange can be further obtained through these addresses, helping users understand the real Bitcoin holdings of the exchange. Several experiments are conducted to evaluate the effectiveness of the proposed Bitcoin address tagging method. Our method has very little dependence on off-chain information. Only one address is needed for each exchange as a seed to find out all the other addresses. Such a seed address can be easily obtained by depositing some Bitcoin into the exchange or withdrawing some from it, which makes our method feasible for all exchanges.
Due to the decentralization, irreversibility, and traceability, blockchain has attracted significant attention and has been deployed in many critical industries such as banking and logistics. However, the micro-architecture characteristics of blockchain programs still remain unclear. What's worse, the large number of micro-architecture events make understanding the characteristics extremely difficult. We even lack a systematic approach to identify the important events to focus on. In this paper, we propose a novel benchmarking methodology dubbed BBS to characterize blockchain programs at micro-architecture level. The key is to leverage fuzzy set theory to identify important micro-architecture events after the significance of them is quantified by a machine learning based approach. The important events for single programs are employed to characterize the programs while the common important events for multiple programs form an importance vector which is used to measure the similarity between benchmarks. We leverage BBS to characterize seven and six benchmarks from Blockbench and Caliper, respectively. The results show that BBS can reveal interesting findings. Moreover, by leveraging the importance characterization results, we improve that the transaction throughput of Smallbank from Fabric by 70% while reduce the transaction latency by 55%. In addition, we find that three of seven and two of six benchmarks from Blockbench and Caliper are redundant, respectively.
Today we have enormous amount of data available in every sector, with the advent of technology available, it is possible to provide solutions to many problems. In this paper we are going to provide solutions to the problems related to healthcare data management using Machine Learning and Blockchain. Extracting only the relevant information from the data is possible with the use of Machine Learning. This is done using trained algorithms. Once this data is stored, the next problem is Data sharing and its reliability. This is where Blockchain comes into picture. The consensus in Blockchain technology makes sure that data is legitimate and transactions are secure. Blockchain technology can potentially change health care management for the better by placing patient at the epicentre of the healthcare system and increasing the privacy and interoperability of health data. This paper focuses primarily on solving healthcare data management problems by using Blockchain technology and including some indispensable features using Machine Learning.
G. A. Pierro, Henrique Rocha, Roberto Tonelli, Sté́phane Ducasse
The Ethereum Blockchain is a distributed database that records all transactions and smart-contracts created on the platform. In Ethereum blockchain, the user needs to set a Gas price to get a transaction recorded. To have the transaction recorded, the Gas price has to be greater than or equal to the lowest Ethereum transaction fees. To help the users and smart contracts to set the right Gas price, the Gas Oracle categorizes the gas price into categories based on the interval of time the user might be willing to wait and for each of them suggests a gas price to set. The paper aims to verify the hypothesis that the predictions made by the EtherGasStation Oracle have a margin of error greater than the margin of error declared by it (2 %). We collected data in two-months time from the EthGasStation Oracle which predict the Gas Price every time that 100 blocks are added to the Ethereum Blockchain. In the same time frame, two-months, we also collected over 10 million transactions from a Transaction Pool. By cross-checking the data collected by the Transaction Pool and the Gas Oracle, the study revealed that the Gas Oracle fails more often than it advertises.
Crowdfunding is a form of fundraising in which we collect small amounts of money from a large number of people, which totals to a large amount. In short it is the pooling of large funds which are required for any projects or ideas. Over the years there has been a decrease in overall investment and had remained a complicated and a troubled domain due to lack of transparency and control by a central authorities being some of the primary reasons. There have been many attempts on solving this issue but none have been completely successful. Blockchain is a technology through which we can successfully achieve transparency in transactions up to some extent.
Cryptocurrency merchandising is growing as an attractive area of investment. Bitcoin is much preferred over other cryptocurrencies in the world and hence is becoming more popular. But, the bitcoin price is extremely volatile. So, the forecasting of its price is highly desirable. As nature inspired-machine learning is being used extensively for time series analysis and prediction, it can be explored for bitcoin prediction as well. Also, as bitcoin is gradually increasing as a promising virtual asset, its volatility needs to be measured. This paper unveils the consequence of using ChebyShev Ploynomial Neural Networks (CHPNN) for Bitcoin pricing process. The evolutionary algorithms: Particle Swarm Optimization (PSO) and Differential Evolution (DE) are utilized for training the model. This study analyses the performance of the model through three different error measures: Root Mean Square Error (RMSE), RRSE (Relative Root Square Error) and SSE (Sum of Squares Error). It shows that DE-CHPNN predicts better day-ahead price of bitcoin.
The smart factory is a representative element reshaping conventional computer-aided industry to data-driven smart industry, while it is nontrivial to achieve cost effectiveness, reliability, mobility, and scalability of smart industrial systems. Data-driven industrial systems mainly rely on sensory data collected from statically deployed sensors. However, the spatial coverage of industrial sensor networks is constrained due to the high deployment and maintenance cost. Recently, mobile crowd sensing (MCS) has become a new sensing paradigm owing to its merits, such as cost effectiveness, mobility, and scalability. Nevertheless, traditional MCS systems are vulnerable to malicious attacks and single point of failure due to the centralized architecture. To this end, in this article we integrate MCS with industrial systems without introducing any additional dedicated devices. To overcome the drawbacks of traditional MCS systems, we propose a blockchain-based MCS system (BMCS). In particular, we exploit miners to verify the sensory data and design a dynamic reward ranking incentive mechanism to mitigate the imbalance of multiple sensing tasks. Meanwhile, we also develop a sensory data quality detection scheme to identify and mitigate the data anomaly. We implement a prototype of the BMCS on top of Ethereum and conduct extensive experiments on a realistic factory workroom. Both experimental results and security analysis demonstrate that the BMCS can secure industrial systems and improve the system reliability.
To make good use of valuable Internet of Things (IoT) data assets, this paper proposes a trust-aware IoT data economic system (TIDES) with complete IoT data pricing, trading and protection functions. To ensure reliable and automatic data trading, the entire trading process is automatically performed by smart contracts on a hierarchical blockchain. Moreover, we develop several sophisticated methods to ensure the efficiency and service quality of TIDES. First, a complete evaluation model that takes the data trading profile and reputation into consideration is proposed for both suppliers and demanders to assess the trustworthiness of their trading partners. Second, a client-centric data value evaluation model and a game-theory-based pricing model are used to promote win-win transactions in which the demanders obtain higher quality data at an acceptable price and the suppliers receive higher profits. Third, a dispute arbitration model is invoked to detect suspicious trading and refund these payments automatically. TIDES further utilizes a multi-access edge computing (MEC) architecture to alleviate the huge burdens of IoT devices from blockchain operations, reduce the trading latency, and help mobile devices to trade IoT data. The simulation results have shown the advantages of TIDES in terms of trading time, storage overhead, data trading profit, quality data trading, pricing efficiency, and reliability on data asset management and trading.
Bitcoin is a virtual and decentralized cryptocurrency that operates in a peer-to-peer network providing a private payment mechanism. It is a multi-billion dollar cryptocurrency, and hundreds of other cryptocurrencies are created based on it. Bitcoin is based on Open-Source (OSS) software development, and OSS is a convenient way to qualitatively measure software development and growth. This thesis presents the first comprehensive study of the Bitcoin ecosystem in GitHub organized around 481 most popular and actively developed Bitcoin related projects over eight years (2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018).
In Ethereum, reaching a transaction consensus costs a certain number of gases, which should be purchased by users in their self-defined gas prices. Generally, the higher the gas price, the shorter the time is spent on reaching consensus. Since the transaction gas prices still vary greatly in a block, generating a reasonable price that can make a trade-off between the consensus time and the gases cost is of great significance. In this paper, we propose a Machine Learning Regression-based gas price predicting approach (MLR), aiming to find the lowest transaction gas price in the next block for carrying out economical Ethereum transaction. Specifically, we identify five influencing factors (i.e., difficulty, block gas limit, transaction gas limit, ether price, and miner reward) from the Ethereum transacting process and resort the classic machine learning regression to build the predicting model. Our empirical study on 194,331 blocks implies that the proposed MLR approach works well and can save $17,552.2 for all transactions in the 74.9% accuracy.
In the Ethereum network, miners are incentivized to include transactions in a block depending on the gas price specified by the sender. The sender of a transaction therefore faces a trade-off between timely inclusion and cost of his transaction. Existing recommendation mechanisms aggregate recent gas price data on a per-block basis to suggest a gas price. We perform an empirical analysis of historic block data to motivate the use of a predictive model for gas price recommendation. Subsequently, we propose a novel mechanism that combines a deep-learning based price forecasting model as well as an algorithm parameterized by a user-specific urgency value to recommend gas prices. In a comprehensive evaluation on real-world data, we show that our approach results on average in costs savings of more than 50% while only incurring an inclusion delay of 1.3 blocks, when compared to the gas price recommendation mechanism of the most widely used Ethereum client.
In this work we analyze the blockchain forking events, blockchain partitioning, and duration of inconsistent state of the ledger in a Bitcoin delivery network. Using a comprehensive probabilistic model, we obtain the probability distribution of two- and three-way forks, the forked partition sizes, and the duration of ledger inconsistency until the resolution. We show that the three-way forking probability is substantially lower than that of a two-way forking and that the partition sizes in the case of two-way forking tend to equalize when the number of nodes increases. Finally, we show that the duration of ledger inconsistency state exhibits long tail probability distribution which means that successive forking events can force the ledger to remain inconsistent for long time.