Hi everyone! Since I am required to collect and correctly classify some KPIS on a monthly basis, I need support automating this process using machine learning. Every month I receive an excel file containing a plenty of rows. Each row corresponds to an order placed by a customer. The columns, such as invoice description, order description and others, instead contain descriptive textual information related to each order. This information is currently read in order to fill in the last column, where I manually categorize each order to the product/service root to which the order belongs. This last task must be done accurately, given that even a single keyword in a single column among all can allow to correctly classify the root of each order. I want to develop a machine learning model that automatically fills the classification column on the new excel files I will receive along the next months. I hypothesized that he could learn to do it by identifying patterns and correlations between the classifications and keywords made in the previous files. Briefly, I would like him to follow the same reasoning that I followed in classifying each order. I kindly ask if anyone can advise me which is the best approach, the best ML method and which nodes are needed to achieve this goal. Thanks in advance :slight_smile:

Knime Machine Learning to automate KPIs collection

iCFO February 4, 2023, 1:59am 6

Here is a forum post that has a workflow and detailed explanation of text processing and classification. I haven’t had time to dive in myself yet, but it looks intriguing.

1 Like