Complaints Classification using NMF

Topic Modeling using Non-Negative Matrix Factorisation

Here we will be using NMF(Non-Negative Matrix Factorisation) with the help of this technique, which is an approach under topic modelling, we will detect patterns and recurring words present in each ticket. This can be then used to understand the important features for each cluster of categories. By segregating the clusters, we will be able to identify the topics of the customer complaints.

What is NMF(Non-Negative Matrix Factorisation) ?

Non-Negative Matrix Factorisation is a statistical method that helps us to reduce the dimension of the input corpora or corpora. Internally, it uses the factor analysis method to give comparatively less weightage to the words that are having less coherence. Non-Negative Matrix Factorisation (NMF) is an unsupervised technique so there are no labeling of topics that the model will be trained on. The way it works is that, NMF decomposes (or factorises) high-dimensional vectors into a lower-dimensional representation. These lower-dimensional vectors are non-negative which also means their coefficients are non-negative.

Here we have customer complaints, which are unstructured in nature. So, traditionally, companies need to allocate the task of evaluating and assigning each ticket to the relevant department to multiple support employees. This becomes tedious as the company grows and has a large customer base.