- Model -----> Outcome
- Forcasting Model -------> Forecast
- Model 1
- Model 2 ---> Ensemble( A collection of things that intended to be used together) -->
Forecast
- Model 3
- Understand financial consequences
Gross Sales before cancellation
|Cancellation Rate Value
|Return Rate Value
||=>
Net Merchandise Volume
= Sequence of data points in time order recorded in equally distanced time period
- Trend
- Seasonality
- Error
- Airbnb missed the earning expectations
- Airbnb struggling in the US market
Forcasting Product
- Demand in Newyork
What data we have
- Holidays, Temperature and Marketing Investment
- Daily Demand($)
- Historical Data to find patterns
- Predict demand of the incoming month
if we are considering the seasonality in terms of percentage, then we should consider multiplicative seasonality if we are dealing with absolute values then it's additive
Additive Trend + Seasonality + Error
Multiplicative Trend * Seasonality * Error
Auto- Correlation plorts (ACF)
- A check to see if there is information in the past to make a decision
- correlation with past
- Slope - Linear regression for a year
DataAnalysis : DataAnalysis.ipynb