Classifying Texts for an Organization - Machine Learning
Business Case:
On average, technicians across the country send out 9,000 work orders daily to activate contracted services.
These operations can experience various types of blocks, which, depending on the code, require different handling procedures.
The audit process, based on the technician's explanation justifying the code used, is 100% manual.
A Machine Learning model can not only automate the audit to detect deviations early but also help orchestrate work orders more effectively, directing them to the correct processing channels and preventing repeated outgoings that lead to non-compliance.
Models:
Three models will be developed to evaluate the behavior and predictive effectiveness:
Neural Network: Recurrent neural networks (RNNs) are a type of deep neural network particularly well-suited for processing data sequences, such as text. RNNs can remember information from previous inputs, allowing them to understand the context of a sequence, making them ideal for working with natural text.
KNN: This is a good option for classification problems with small to medium-sized datasets where interpretability is important. It is based on finding the K data points closest to the new data point and assigning it the majority class among those K neighbors.
XGBOOST: This is a boosted decision tree algorithm that combines multiple weak decision trees to create a stronger model. XGBoost includes regularization techniques to avoid overfitting.
Presented by: Bruno Conti. I hold a degree in Information Systems from the University of Salvador. I am a curious professional, always seeking the best possible solutions. I am passionate about building innovative and agile teams.