NLP Ise 2

Research Paper 1: "AI-based chatbots in customer service and their effects on user
compliance" by Martin Adam, Michael Wessel, and Alexander Benlian
Introduction
The paper focuses on the use of AI-based chatbots in customer service, examining their impact
on user compliance. The advent of live chat interfaces in e-commerce has led to a shift from
human service agents to conversational software agents like chatbots, which are often based on
artificial intelligence (AI). Despite the cost and time efficiencies, AI-based chatbots sometimes
fail to meet customer expectations, affecting user compliance with chatbot requests. The study
leverages social response theory and commitment-consistency theory to empirically analyze
how verbal anthropomorphic design cues and the foot-in-the-door technique influence user
request compliance in self-service interactions.
NLP Aspects and Trends
The study leverages advanced NLP capabilities to enable chatbots to engage in more
human-like interactions through the use of verbal ADCs. These capabilities include
understanding user inputs, processing emotions, and generating empathetic responses, thereby
enhancing the user's perception of social presence and ultimately their willingness to comply
with chatbot requests.
Methodologies and Technologies
1. Experimental Design: A randomized online experiment was conducted in the context of a

customer-service chatbot for online banking. The experiment involved a chatbot interaction with
predefined dialogue flows and a messenger interface.
2. Anthropomorphic Design Cues (ADCs): Chatbots in high ADCs manipulation conditions

exhibited human-like traits such as identity, small talk, and empathy to enhance user interaction
and perception of the chatbot.
3. Foot-in-the-Door (FITD) Technique: This involved making a small initial request followed by
a larger request to influence user compliance.
4. Data Analysis: Binary logistic regression analysis and mediation/moderation analyses were
employed to examine the direct and indirect effects of ADCs and FITD on user compliance.
Applications
This research is highly relevant to the field of e-commerce and customer service, where
AI-based chatbots are increasingly employed. Understanding the factors that influence user
compliance can enhance the effectiveness of chatbots in various settings, including banking,
retail, and customer support.
Drawbacks
1. Experimentation Limitation: The study was confined to an experimental setting, which may
not fully replicate real-world chatbot interactions.
2. Cultural Limitations: The sample consisted of German participants, which might not
represent global user behavior.
Additional Insights
- Social Response Theory and Anthropomorphism: The study showed that anthropomorphic
design cues in chatbots, such as identity and empathy, can increase user compliance by
enhancing the chatbot's perceived social presence.
- Commitment-Consistency Theory: The FITD technique's effectiveness was attributed to the
user's desire for self-consistency, irrespective of the chatbot's social presence.
Conclusion
The study provides significant insights into the design and application of AI-based chatbots in
customer service. It highlights the importance of NLP-driven verbal anthropomorphic design
cues and psychological compliance techniques in enhancing user interactions with chatbots.
While promising in increasing user compliance, these strategies must be carefully implemented
to ensure authenticity and effectiveness in various service contexts.
The findings advocate for the strategic design of chatbot interactions to foster user compliance,
highlighting the evolving role of NLP in creating more effective, human-like chatbot interactions
References:
https://link.springer.com/article/10.1007/s12525-020-00414-7
Research Paper 2: Chatbot: An automated conversation system for the
educational domain
Introduction
The integration of chatbots in the educational domain signifies a revolutionary approach to

facilitating accessible, efficient, and interactive learning. This report examines the development
of an educational chatbot designed to address questions from users within this domain.
Employing Random Forest as an ensemble learning method and focusing on Natural Language
Processing (NLP) techniques, the chatbot aims to provide accurate, contextually relevant
responses to enhance the educational experience.
Methodologies and Technologies
1.Data Collection and Preprocessing
The foundation of the chatbot's capabilities is a dataset comprising approximately 1500

educational dialogues, meticulously collected from an unspecified educational organization. This
dataset underwent extensive preprocessing, including data cleaning, integration, transformation,
reduction, and discretization, to transform unstructured data into a structured format suitable for
further analysis and feature extraction. Python scripts played a crucial role in this phase, setting
the stage for the NLP-centric feature extraction process.
2. Feature Extraction and NLP Emphasis
The core of the chatbot's intelligence lies in its NLP-driven feature extraction process. Utilizing
NLTK (Natural Language Toolkit) and WordNet, the system identifies and extracts a variety of
features from the text, including question types (e.g., "why," "what," "where"), nouns, verbs,
Term Frequency (TF), Inverse Document Frequency (IDF), and TF-IDF values. These features
are instrumental in understanding the semantics and context of user queries, allowing for more
accurate and relevant responses. The employment of WordNet significantly enhances the
chatbot's lexical database, enabling it to identify synonyms and hyponyms, thereby enriching its
Bag-of-Words (BoW) model and improving its generative capabilities.
3. Model Building: Random Forest and NLP Integration
The application of the Random Forest algorithm facilitates the classification of responses into
retrieval-based or generative types. This algorithm, known for its robustness against overfitting,
combines multiple decision trees to make more accurate predictions. In this chatbot, the
integration of Random Forest with NLP features enables the system to handle the complexity of
natural language, understand user intents, and generate responses that are not only
contextually accurate but also linguistically coherent.
Applications
This chatbot finds its application primarily in the educational sector, serving as a 24/7 virtual
assistant capable of answering a wide array of educational queries. By leveraging NLP, the
chatbot can interpret the nuances of human language, making it an invaluable tool for students
seeking instant responses to their questions.
Drawbacks and Challenges
While the chatbot demonstrates considerable potential, it faces challenges inherent in NLP and
machine learning. The accuracy and relevance of responses can be limited by the scope of the
training dataset. Additionally, the complexities of human language, including idioms, sarcasm,
and context-dependent meanings, pose significant challenges to understanding and generating
responses. The reliance on a predefined dataset also means the chatbot may struggle with
queries outside its training scope.
Future Directions and NLP Enhancements
Future improvements should focus on expanding the chatbot's dataset and enhancing its NLP
capabilities. Incorporating advanced NLP techniques such as deep learning models for
language understanding and generation could significantly improve the chatbot's ability to
understand complex queries and generate more coherent and contextually relevant responses.
Additionally, ongoing training with a more diverse dataset could enable the chatbot to handle a
broader range of topics and question types.
Conclusion
The development of an educational chatbot utilizing NLP and Random Forest represents a
significant advancement in digital learning tools. By addressing the unique challenges of
processing natural language, this chatbot offers a promising approach to enhancing the
educational experience. Future enhancements in NLP technologies and methodologies hold the
potential to further refine and expand the capabilities of such educational chatbots, making them
even more versatile and effective tools for learners worldwide.
Reference
A. Mondal, M. Dey, D. Das, S. Nagpal and K. Garda, "Chatbot: An automated conversation

system for the educational domain," 2018 International Joint Symposium on Artificial Intelligence
and Natural Language Processing (iSAI-NLP), Pattaya, Thailand, 2018, pp. 1-5, doi:
10.1109/iSAI-NLP.2018.8692927. keywords: {Chatbot;Feature extraction;Random
forests;Vegetation;Decision trees;Training data;Organizations;Chatbot;Educational
Domain;Question answering;Machine Learning},
URL: https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8692927&isnumber=8692798

NLP Ise 2

Uploaded by

Copyright:

Available Formats

NLP Ise 2

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

NLP Ise 2

Uploaded by

Copyright:

Available Formats

Research Paper 1: "AI-based chatbots in customer service and their effects on user

compliance" by Martin Adam, Michael Wessel, and Alexander Benlian

NLP Aspects and Trends

Methodologies and Technologies

1. Experimental Design: A randomized online experiment was conducted in the context of a

2. Anthropomorphic Design Cues (ADCs): Chatbots in high ADCs manipulation conditions

The integration of chatbots in the educational domain signifies a revolutionary approach to

Methodologies and Technologies

1.Data Collection and Preprocessing

The foundation of the chatbot's capabilities is a dataset comprising approximately 1500

2. Feature Extraction and NLP Emphasis

3. Model Building: Random Forest and NLP Integration

Drawbacks and Challenges

Future Directions and NLP Enhancements

A. Mondal, M. Dey, D. Das, S. Nagpal and K. Garda, "Chatbot: An automated conversation

You might also like