🤖 This project is showing the classification model tests and compared their accuracy with confusion matrices. Besides, the regression model was used for revenue prediction by each category of products.

• About
• Product Details
• SQL_Join
• Accuracy for Classification model
• Decision Tree Regression
• Conclusion

Roventure is a store selling sports items. However, it involved some loss due to overstock, and the store doesn’t know how to adjust the sales price to reduce the overstock items. To control the inventory, reduce the overstock issues, and keep the competitive pricing. I need to use the regression model to predict the price based on the parameters: selling quantities, return items, cost, and the categories.

Power BI, SQL, Python or Jupyter Notebook

Used Power BI to connect each table for monthly details:

Used SQL to join tables and get the cost and revenue by aggregation for 2017 sales:

Read document from SQL table and checked the category:

Used 6 Classification models for accuracy test, and DT had the highest accuracy since there were 71 mistakes.

1. Logistic Regression
2. Decision Tree(DT)
3. K neighbors classifier (KNN)
4. Linear Discrininant Analysis (LDA)
5. Gaussian Naive Bayes(GNB)
6. SVC

Used DT Regression for revenue prediction:

a. Tested accuracy: the errors were very low, which meant the model was good. b. For model categores: 1--> Bikes 2--> Components (No sales in 2017） 3--> Clothing 4--> Accessories c. for example: if category is 3, qty is 2, cost is 4.5, price is 6.00, revenue would be 6.25 . shape(1,4)--> 1 observation and 4 features you need to input.

In conclusion, it is very important to figure out the label and features in classification. Also, testing a model's accuracy is necessary before applying it to a model for prediction. This prediction can help set up the selling price for increasing revenue.