3. Homework 1: COVID-19 Cases Prediction (Regression) 实验目的: 熟悉PyTorch。 了解基本的DNN培训技巧。 使用深度神经网络(DNN)解决回归问题。 下载数据集: !gdown --id '1kLSW_-cW2Huj7bh84YTdimGBOJaODiOS' --output covid.train.csv !gdown --id '1iiI5qROrAhZn-o4FPqsE97bMzDEFvIdg' --output...
3、继续通过特征选择来调整结果,经过大量尝试,发现使用前3天的COVID-like illness 和前2天的Tested Positive Cases 特征效果最佳,并添加归一化使得训练过程更加稳定,得到了接近Boss Baseline的结果。 在这里插入图片描述 4、接着通过大量调参工作:学习率、优化器、Batch、EarlyStop、Model修改都没得到更好的验证集分...
States(40)及各兩天的 COVID-like illness (4)、Behavior Indicators (8)、Mental Health Indicators (5)、Tested Positive Cases (1)我們要透過DNN,輸入Day1, Day2以及Day3的資料,來預測第三天的tested_positive。 (2) Feature Selection 由於Features都是以numerical的形式表示,可以先初步應用correlation來檢驗...
公开项目>COVID-19 Cases Prediction (Regression) COVID-19 Cases Prediction (Regression) Fork 1 喜欢 0 分享 李宏毅 ML2021Spring - HW1 yyinpu BML Codelab 2.1.2 Python3 初级 2021-10-21 21:38:45 版本内容 数据集 Fork记录 评论(0) 运行一下 regression-v1 2021-10-22 22:32:10 请选择预览...
simple baseline过很简单,每行数据不做任何处理,直接作为网络的输入就行,网络也不用优化。 medium baseline需要加一些优化。 首先前面表示州的one-hot向量过于稀疏,所以我直接忽略。 由于输入数据全是正数,而且我使用了relu,所以训练中有时会导致loss一直不变,所以我将输入数据进行了归一化。
简介:COVID-19 Cases Prediction (Regression) Objectives: Solve a regression problem with deep neural networks (DNN). Understand basic DNN training tips. Familiarize yourself with PyTorch. Task Description COVID-19 Cases Prediction Source: Delphi group @ CMU ...
COVID-19 Cases Prediction Source: Delphi group @ CMU A daily survey since April 2020 via facebook. Try to find out the data and use it to your training is forbidden Given survey results in the past 5 days in a specific state in U.S., then predict the percentage of new tested positiv...
李宏毅2021机器学习HW1 在Google Colab中运行成功,未在PP中运行 - 飞桨AI Studio
In this section, we summarize the previous studies in the context of COVID-19 time series prediction. Since the publicly available data of COVID-19 contains daily statistics of the confirmed cases, so it is considered as a time series data and the time series forecasting techniques can be ex...
. Our best model also beats the Ensemble model on average in three- and four-week ahead prediction horizons. As part of our evaluations, we also compare the predictive power of Facebook-derived and SafeGraph-derived features in the context of our models for predicting new COVID-19 cases....