Skip to content

Latest commit

 

History

History
34 lines (27 loc) · 917 Bytes

README.md

File metadata and controls

34 lines (27 loc) · 917 Bytes
Raw

Time_Series_Analysis_Forecast

Time Series Analysis for House Prices -- various methods: ARIMA variations, RNN and Prophet

Table of Contents

  • Project Aim
  • Dataset
  • Libraries
  • Models Selection

1. Project Aim

Time Series Analysis Project using different statistical and machine learning tools to descibe and forecast time-series data.

2. Dataset

The dataset ma_lga_12345.csv. It is the sales data of houses and flats in quarterly periods. The sale data in column MA.

3. Libraries

  • Arima and Sarimax: numpy, pandas, statsmodels, pmdarima

  • RNN: numpy, pandas, keras, statsmodels

  • Prophet: to be added

4. Models

  • The development of the ARMA, ARIMA model: differencing, checking for non-stationary data, auto-correlation, partial auto-correlation, seasonality, forecasting
  • Development of sequence RNN models: to be added
  • Development of forecasting with FB Prophet library: to be added