Regression and Univariate Models

Powerful models for predictive analytics

We mainly use supervised models for our regression analysis. This includes multiple linear regression, and logistic regression for classification.

Univariate models only have a single output. We use auto regressive models and moving average models for basic trends, however, to capture more complex patterns we incorporate SARIMA and SARIMAX. Both these models can capture seasonality trends, and SARIMAX can have multiple variables as inputs. We use these models for timeseries forecasting.

Neural Networks

LSTM Models are Remarkably Accurate

We create LSTM models for more accurate timeseries forecasting. These models are in the majority of cases better than regression, autoregressive, and moving average models. LSTM stands for long short-term memory, and they are the superior model in capturing data trends far back in history.