Oracle Analytics Cloud and Server Idea Lab

Welcome to the Oracle Analytics Community: Please complete your User Profile and upload your Profile Picture

OAC DV - Add linear Regression functionality

Needs Votes
19
Views
5
Comments

Added linear regressions in the Analysis section, then forecast for certain visualizations such as curves (see the example on the right). In addition to the forecasts already present such as ARIMA, SARIMA, ETS, etc.

image.png
4
4 votes

Needs Votes · Last Updated

Comments

  • Brendan T
    Brendan T Rank 6 - Analytics Lead

    Yes this would be useful and some additional regression lines in other visualizations would also be helpful, along with an AB line

  • Thanks @Thomas HYRIEN and @Brendan T

    currently using right-click trendline, you can add a linear(regression)/poly/expo trendline on the viz directly, see below, that should already help with the initial question. As for slope and intercept (was that your question ?) these will are provided when the trendline is built on a scatter plot. But when the trendline is using date, X is then normalized to a sequential number and A&B become conventions (funtionally meaningless). Let me know if this does not answer your question

    image.png

    Thanks

  • Thomas HYRIEN
    Thomas HYRIEN Rank 4 - Community Specialist

    Hi,

    Thanks,

    We would like some additional forecast models in the list in addition to ARIMA, SARIMA, ETS, etc.

  • Brendan T
    Brendan T Rank 6 - Analytics Lead

    While additional algorithms are being requested there is a fuller list of what could be added (some already there)

    Autoregression (AR): Predicts future values based on past values of the same series. 

    • Moving Average (MA): Uses the average of previous data points to smooth out fluctuations and identify trends. 
    • ARIMA (Autoregressive Integrated Moving Average): A popular forecasting model that combines AR and MA concepts, often used for non-stationary time series after differencing to make them stationary. 
    • SARIMA (Seasonal ARIMA): An extension of ARIMA designed to handle data with clear seasonal components. 
    • Exponential Smoothing: A family of methods that give more weight to recent observations, including methods like Holt-Winters that capture trend and seasonality. 
    • Vector Autoregression (VAR): An extension of AR for modeling multiple related time series.
    • LSTM (Long Short-Term Memory) Neural Networks: A type of recurrent neural network well-suited for capturing complex temporal dependencies in time series data. 
    • Prophet: Developed by Facebook, this is a robust forecasting tool designed to handle data with strong seasonal effects and missing values, and is also effective with holiday events. 
    • XGBoost: An efficient and powerful gradient boosting algorithm that can be adapted for time series forecasting on tabular data.