Text Analysis And Classification Using Topic Modeling

A large amount of data is generated and collected daily to understand the trends. Searching and analysing the data becomes complex as more information becomes available. Therefore, to search through, sort, analyse, and comprehend huge amounts of data, we need sophisticated methods and tools. The topic modeling tools that can help organize, comprehend, and summarize a significant amount of textual information are examined in this work. With topic modeling, it is possible to organize, search, and summarize texts by annotating documents with hidden patterns that are present throughout the textual data. This paper uses the Latent Dirichlet Allocation (LDA) in topic modeling for text analysis and classification of data extracted from medical equipment installation logs. When a healthcare product-based company installs their product in a hospital they pay the engineer based on their hourly service. It becomes a crucial concern if these service hours are hiked. To understand the reason for the service hour hike, an analysis is required. This paper aims in classifying the product installation text, entered by field personnel based on their daily activities until the completion of the installation in the high/medium/lowcost category. After classification one can analyse the text that belongs to the high-cost category only to understand the challenges faced during field installation. The analysing word space is drastically reduced while analysing the text by application of LDA.