第一部分：

感谢我partner的精彩演讲。 - "Thanks for my partner's fantastic speech."
下面由我给大家进行Demonstration 部分的展示。 - "Next, it's my turn to give a demonstration of the content for the 'Demonstration' section."

我们首先分别从每日，每月，和每年的回报率着手，来了解泰达币的趋势。 - "First, we'll dive into the daily, monthly, and annual returns to understand the trends of Tether (USDT)."
在这部分我们主要关注的有三个方面， Volatility，Trend Analysis，Market Events。 - "In this section, our primary focus will be on three aspects: Volatility, Trend Analysis, and Market Events."

在每日回报率中，我们可以看到无论是波动多么的大，例如在2020年的1月波动很大，但是在2022年波动很小，但是对于回报率来说，他是一种动态平衡的状态。 - "In the daily returns, we can observe that regardless of the extent of volatility, for instance, in January 2020 where the volatility was high, and in 2022 where it was low, the returns exhibit a dynamic equilibrium."

虽然比起2019年，2021年之后的回报率趋势很小，对于普通的玩家来说这并不是一件坏事，所以我们明白并不是稳定币就没有投资的意义。 - "Though after 2019, the trend in returns has been relatively small, it's not necessarily a bad thing for the average investor. This indicates that stablecoins still hold investment potential."

不难看出，结合图表查阅了相关的历史信息后，我们发现在2020年1月开始，USDt的波动异常的大，原因很大一部分是因为在2020年的事件引发的。 - "Upon examining the charts and historical information, we find that the significant volatility in USDT starting from January 2020 was largely due to certain events in that year."

每月的回报率中，一方面对我们每日回报率总结的验证，另一方面也直观的反映了，在每年的12月和1月市场的波动都相对于其他月份大一些。 - "Moving on to the monthly returns, it serves as both a validation of our daily returns summary and a visual representation of how market volatility tends to be somewhat higher in December and January compared to other months."

从年回报率图中，更加虽然2020年之后波动很小，但是他的回报率曲线呈现出一个正弦波的趋势。这里引起了我们的注意，截止到数据最后的日期，这个趋势会继续上升还是下降呢？ - "Now, when we look at the annual returns chart, we see that even though the volatility has been low after 2020, the return curve exhibits a sinusoidal trend. This raises the question: will this trend continue to rise or fall as of the latest data?"

传统的数据分析就无法解决这样的问题，这就需要我们使用机器学习的方式来挑战这一问题。 - "Traditional data analysis cannot address this issue, which is why we need to use machine learning to tackle this challenge."

第二部分

使用机器学习首先面临的挑战就是选取什么样的算法，
The first challenge in using machine learning is selecting the right algorithm.
我们查阅了很多资料和论文，
We've searched through a lot of materials and papers,
但是无论是在网上还是论文里，
but we couldn't find any methods for this issue, whether online or in the papers.
都没有找到关于这个问题的方法。
We couldn't find any methods for this issue.
用了几个方法结果都很差，例如KNN,SVM，
We tried a few methods, like KNN and SVM,
研究一度陷入困境。
but the results were poor, and our research hit a dead end.
但是，本学期的另外一个课启发了我，Neuroinformatics，
However, another course this semester, Neuroinformatics, inspired me,
回报率的曲线正如脑电波一样，
as the return rate curve is much like brainwave data,
是时间序列并且有正负值，
it's a time series with positive and negative values,
最后，选取了在该学科常用的机器学习方法来进行分析。
so I decided to use the machine learning methods commonly used in this field.
正如现在展示的结果一样，
As you can see from the results presented now,
预测的结果和测试集趋势高度相关，
our predictions are highly correlated with the test set's trends,
很好的预测了回报率的趋势。
effectively forecasting the return rate trends.
这就是我们本次研究的所有内容，
This concludes our research for this project,
虽然本次作业结束了，
even though the assignment is over,
但是并不意味着研究结束，
it doesn't mean the research ends,
我希望我们的报告可以给进行区块链数据分析的研究人员一些新的思路，
I hope our report can provide some fresh insights to researchers working on blockchain data analysis,
更好的发展区块链生态。
and contribute to the further development of the blockchain ecosystem.