Abstract :

Unstructured Customer Communication Such As Helpdesk Transcripts, Support Tickets, And Chat Logs Frequently Contains References To Geographic Locations That Are Valuable For Routing, Analytics, And Service-quality Monitoring, Yet These Mentions Are Easy To Overlook When Reviewed Manually At Scale. This Work Presents A Lightweight Naturallanguage-processing Pipeline That Automatically Identifies Place Names Embedded Within Free-form English Sentences And Exposes The Pipeline Through A Web-based Portal For Interactive Use. Input Text Is Passed Through A Preprocessing Stage That Performs Sentence And Word Tokenization, Part-ofspeech Tagging, Stop-word Removal, And A Comparative Stemming/lemmatization Step, After Which Token-level Features Derived From POS Tags And Surrounding Context Are Used To Train Classical Machine-learning Classifiers, Namely Logistic Regression, Multinomial Naive Bayes, And Support Vector Machine, For Place-name Recognition. A Parallel Helpdesk-call Sentiment Classifier Is Trained On The Same Preprocessing Backbone To Demonstrate The Pipelines Reuse Across Tasks. The System Is Delivered As A Flask-based Portal With Separate User And Administrator Views: Users Submit Sentences And Receive The Identified Place Names Highlighted Inline Together With A Breakdown Of The Preprocessing Steps Applied, While Administrators Can Re-run The End-to-end Training Pipeline, Inspect Dataset Cleaning Statistics, And Compare Model Accuracy Through Generated Charts. Evaluation On A Custom-labelled Placename Dataset Shows That The Logistic Regression And Support Vector Machine Models Achieve The Strongest Recognition Accuracy, Indicating That A Carefully Engineered, Classical Feature-based Approach Remains A Practical And Interpretable Option For Domain-specific Place-name Identification.

Published:

01-8-2026

Issue:

Vol. 26 No. 8 (2026)

Page Nos:

1146 - 1153

Section:

Articles

License:

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

How to Cite