Trial Business Analyst Time Series Modeling Assignment

Assignment Task

Important Points

(1) Attachment: Raw Excel Data.

(2) Main Objective: Predict the Domestic Market (Contract) Blow Molding, Low price.

(3) Build a univariate ARIMA, LSTM, and Lazy Predict with thoughtful features from the list. If you can add any other models that would be great. For the final model please highlight period where the confidence level went down for historic forecasts.

(a) Use a combination of models of your choice to benchmark the accuracy, you can use both classification (up or down), or time series models. Our preferred models are MLP, RNN, LSTM, or GRU - for time series, and Logistic, Random Forest, Naive Bayes Classifier for classification. You are free to use other models.

(4) Evaluation table. Show RMSE, MSE and R2, and also convert directional accuracy measure. Present a comprehensive model reliability matrix with error and classification metrics.

(a) final model performance benchmarking table, show the results and performance metrics for all your models for t+1, 2, 3, etc.. Determine which model is better or worse based on RMSE, MSE, Directional accuracy, etc.

(5) Last but not least use the features space, correlation or other analysis (like XGBoost) to highlight from point one which feature space indicator might help explain from point 1 time periods when confidence level went down historically. Provide a report with a succinct visualization of results and all your different scripts (class object-oriented good scripting practices) and final Python notebook.

This Computer Science has been solved by our PHD Experts at My Uni Paper.