Creating a Predictive model in Knyme Analytics Assignments | Online Assignments Services

you will carry out a project in which you will have to provide a solution to a specific problem based on a case . You will be expected to choose an appropriate artificial intelligence model to make predictions and provide one evaluation strategy to assess the performance of the model.

The predictive mode must be developed step by step:
– The data must be clustered (k-means, k-medoids…) based on total sales.
– At least 2 different predictive models for sales channel need to be tested
(random forest, Bayes…) using the clustered data as an outcome.
– The 2 models must be scored and visualized in a bar chart.
– The report will be based in the model obtaining the highest score.

the report must contain the following elements:
– An explanation of Clusters and Predictive algorithms implemented. Why do they have been selected and what is the outcome:
o Clusters: Explain the clustering method and number of generated clusters and population of each of them.
o Learner and Predictive algorithms: Scores obtained by each model (at least 2).
– Cross validation: the accuracy comparison will be needed for marking.
– Total sales per channel (actual vs. predicted) in a bar chart.
– Data supported actionable recommendations (increase the investment
on traditional channel, increase discounts on line, advice on potential sales decrement…)
– Additional information based on individual cluster analysis will be taken on consideration. Individual recommendation for each segment (in case a cluster by cluster analysis has been done).
– Screenshot of the actual workflow created in Knime (Artificial Intelligence software).