90% accuracy in Sentiment Analysis

Combination of models

At Palowise, we have discovered that using multiple models for sentiment analysis can be advantageous in several ways. Firstly, it allows us to have more control over the types of errors that the model could make. For instance, as we provide a negative alert service, it is crucial that we do not miss any negative mentions. Even though this method does not always guarantee higher accuracy, it can assist us in capturing the most negative mentions by utilizing the predictions from multiple models together. Secondly, combining models can help us identify problematic mentions that cause significant conflict between different prediction models. We can then assign these mentions to our analysts for further examination and classification. Lastly, by using multiple models, we can reduce the reliance of individual models on the amount of training data we have available. As our analysts classify a certain percentage of mentions for each client on a monthly basis, the volume of training data for each project increases gradually. Individual models can perform differently depending on the amount of training data, so combining models can help us overcome this issue to achieve better performance as we have more data to work with.

Deep learning vs Machine learning?

In the era of deep learning, one might wonder why we test classic machine learning algorithms. The answer has to do with data. Deep learning requires large amounts of data to be effective. Classic machine learning models therefore, may be a better choice until enough data is collected for each client. While deep learning has its advantages, classic machine learning algorithms can still be effective in certain situations, particularly when working with smaller datasets.

info: Partners in the development of our Sentiment Analysis technology are the Research Group of the Department of Cultural Technology and Communication at the University of the Aegean and the Laboratory of Knowledge and Uncertainty of the University of Peloponnese, through the Operational Program Competitiveness, Entrepreneurship and Innovation, under the call RESEARCH – CREATE – INNOVATE.