For a relatively simple strategy model such as the autoregressive model, both optimization methods can actually be used. Relatively speaking, the latter method may be closer to the intention of actual trading. However, the optimization method used here is determined to be the former one, which is to obtain coefficient values ​​by fitting an autoregressive model. This is done firstly to show you actual cases of different optimization methods, and secondly because in actual research, if more For more complex prediction models or methods, especially some models with relatively fixed fitting technology such as neural networks, the calculation of the latter optimization method is too complex, so it is rarely used in practice. #ETH #crypto2023 #Binance #Web3 #xiaoyiclub