Enhancing agricultural feedback analysis through Voice User Interface and Deep Learning Integration

Abstract

A substantial amount of critical information in various sectors hinges on consumer feedback, which influences everything from product adoption to overall satisfaction. The agricultural sector, in particular, depends heavily on farmers, whose feedback dictates the success of products and highlights associated challenges. This study aligns with this perspective, emphasizing the importance of understanding farmers’ needs to assist tractor manufacturing industries. However, communication with farmers poses significant challenges, especially when dealing with a large number of farmers across multiple locations. Even in single locations with fewer farmers, information often passes through intermediaries, leading to potential deviations from the original message and creating communication gaps. To address these challenges, we partnered with Dexer, a VUI (Voice User Interface) application-based company. Dexer’s features, including voice recognition accuracy, offline capability, and media support, aids in streamlining the feedback process. This collaboration allows farmers to simply speak into the app to record their feedback, ensuring the information remains raw and unaltered. The partnership aims to enhance communication efficiency and bridge potential gaps in the feedback collection process. In addition to our primary objective, we undertook an exploratory task to analyze various aspects of our dataset. Initially, we used BERT for sentiment analysis on farmer feedback. Later, we compared BERT with RoBERTa, which is specifically designed for sentiment analysis tasks. The primary result we sought from our model was understanding the concrete reasons behind farmers’ ratings. To address this, sentiment analysis was employed to provide reliable information that the responsible parties can use to take informed actions. This study not only bridges the communication gap between farmers and manufacturers but also provides a robust framework for utilizing modern AI techniques to improve product development and adoption in the agricultural sector.