DescriptionAlchemy is a decentralized data model training platform based on chain security encryption operation and random forest training algorithm. Blockchain ecosystem nodes transfer transaction decision data back to beacons to realize data extraction, use, right determination and reward. At the same time, random forest algorithms are integrated to realize decentralized decision model construction and training system, and beacons feedback from data samples are used to modify the fitting model. As a result, a decentralized, self-iterative, autonomous incentive system data model training platform is realized. Alchemy smart contract records data related to transactions / transfers on the platform, and ensures that data ownership and use rights are independent of each other through multi-party secure encryption operations and sandbox data transaction. Alchemy Ecosystem does not store supplier data, but encapsulates the data into sandbox, using only the beacon fitting model fed back by data samples. The Alchemy ecological node can correct the model by transmitting the transaction decision data return beacon and get the corresponding Token reward; the node that mortgages a certain number of tokens can become the verification node, and the verification node can obtain the transaction data training set by the ecological node voting. Alchemy intelligent contract traces all the transaction records through machine learning, processes the data information, takes the maximum ACOIN as the goal, carries on the model construction and training iteration with the random forest algorithm, and obtains the transaction model based on the current training set. When the Alchemy model meets the requirements, the model can generate different trading pairs based on the algorithm and begin to trade automatically. At the same time, the model will be constantly optimized and adjusted.