<![CDATA[Searching for products online is often an inefficient and confusing process, especially when users do not know the exact name of the product or use terms that differ from the search system. Keyword-based searches tend to produce irrelevant results because the system only matches text literally without understanding the meaning. As users increasingly talk to digital devices, voice-based search technology has become a more natural and intuitive alternative. This research aims to develop a semantic-based voicebot supported by Natural Language Processing (NLP) to improve the effectiveness of product searches on e-commerce platforms. The designed system not only recognizes user speech but also understands the context, intent, and semantic meaning of the given commands. The research stages include collecting user voice data, training the Automatic Speech Recognition (ASR) model for voice-to-text conversion, and applying the semantic NLP model for interpreting the context of product searches. The testing was conducted using Indonesian voice commands in a simulated e-commerce scenario. The results showed that the system achieved an average Word Error Rate (WER) of 1.29%, indicating a high level of accuracy in recognizing speech and understanding user intent. The integration between ASR and semantic NLP proved capable of creating a more natural, responsive search experience that resembles the way humans think and communicate when interacting with online search systems.]]>
Copyrights © 2025
