1.1 A study of Customers Insurance Charges
2.1 🛒 LSTM for Time Series Forecasting
Lessons learned: For dealing with times series analysis, it is crucial to understand how dataframe.shift() works. Explanation included in the Notebook!
2.2 😷 Forecasting number of Covid-19 patients using CNN-LSTM Credits to : PseudoLab
- TOO SMALL DATA (only around 300 days !! ) -> Overfitting problem
- 외부요인은 ?
2.3 💍 E-Commerce Jewelry Sales Forecasting
- Notebook contains EDA, Forecasting, Product recommendation.
- for i in range(1, 11):
plt.subplot(5, 2, i)
typ = jewelry_types[i-1]
2.3.1 Customer subscription data
- Autocorrelation analysis is an important step in the Exploratory Data Analysis (EDA) of time series. The autocorrelation analysis helps in detecting hidden patterns and seasonality and in checking for randomness. It is especially important when you intend to use an ARIMA model for forecasting because the autocorrelation analysis helps to identify the AR and MA parameters for the ARIMA model.
-
스마트 팩토리 (시계열 데이터와 머신러닝 프로젝트) : SARIMA를 활용해 외부 데이터 향후 3개월 단위를 예측하고, 2년간의 피처 데이터로 학습된 Random Forest를 갖고 최종 타겟 변수인 생산 수주량 향후 3개월 예측
-
LSTM ...
