A client needs help with a data set project from Kaggle.com
– A detailed report using
1. Introduction: The data you want to examine and which data do you want to forecast
2. Data: (Definition of your data set)
• What is your data?
• what is its source of data, where did you obtain it from (give the link, source, etc).
• what is time span of your data (ex: 2001.1 to 2016.10 daily, quarterly, annual, how many observations)
3. Write the R code how you make R to read the data. (if your data in matrix form, then make sure it will be in time series format. Use “ts” command to create time series data.)
4. Plot your data. Put the graph for the whole period. Write the R code how you create the plot.
5. Dividing the data: Partition the data for 2 periods: Training and Test Period.
write your R codes how you create a Training Period data. (for ex: divide your data into Training and Test period. for example, if you have an monthly inflation data from 2000.1 to 2016.10 as Training Period, and 2016.1 to 2016.9 as Test Period ; then model your data for the Training Period(2001.1 to 2016.1). Then forecast it for the Test Period. See the forecast errors later.
6. Test if there is seasonality and/or trend in the data. Write your R commands.
• If there is no seasonality or trend smooth the data with MA, moving averages.
7. Use Exponential Smoothing (ETS) techniques to model for the Training Period and forecast your data for the Test Period.
• use all the commands in the chapter to see which model fits the best (Holt, Holt-Winters, damped model etc)
• which model do you prefer to use? why?
• write all R commands on the slides then comment on the result.
• forecast for the Test Period.
• what is the forecast performance?
• Decide if you can forecast it for the next horizons?
8. Use ARIMA(p,q) modeling to forecast
• Explain when you can use ARMA modeling. Is the series stationary, for the Trainning Period?
• Put the autocorrelation function of the series(use “Acf” and pacf to see the autocorrelation?) Write your R command. Put the graph on the slide and comment on the autocorrelation function pattern.
• Test if the series is stationary by “adf” command. This is Unit Root Test. Write your R command. What did you decide? is the series stationary or not?
• Use the necessary command to use ARIMA(p,q) to model the series. Write your R command. What do you find? What is the model? What is p and q? Is it Integrated
• Write your ARMA model in equation form. (Be clear about the equation of the model)
• Save the residuals from the ARMA(p,q) model. Write your R command.
• test if residuals are normally distributed? Write your R command.
• plot the acf of the residuals. What do you observe? write your R command in this step.
• forecast the test period,
• what is the forecast error?
9. Compare your forecast error of the both models (ETS and ARMA)? Which model should you use the forecast model for the next horizon?
10. If you run ARIMA(p,q) model for the natural logarithms of the series, did you results change? (Use lambda=0 in your R command and show all your command and result).
Such a cheap price for your free time and healthy sleep
All online transactions are done using all major Credit Cards or Electronic Check through PayPal. These are safe, secure, and efficient online payment methods.