TRAVEL CHAT BOT
A travel chatbot is a type of chatbot that is designed to assist travelers
with various aspects of their journey. It can help travelers find and book
flights, hotels, and activities, provide information about travel
destinations, answer common questions, and provide customer support.
INPUT CAN BE ANY TEXT OR AUDIO
FILE(AUDIO CAN BE CONVERTED INTO TEXT USING
SPEECH RECOGNISATION TECHNOLOGY)
ACTUALLY WE MADE AN
TOURISM PROJECT CAN BE
DEPLOYED IN THAT OR IT
CANE BE DEPLOYED IN
WHATSAPP AND
TELEGRAM
OUTPUT CAN BE TEXT, AUDIO OR ANY
OPTIONS LIKE BUTTONS
BASIC IDEA
BASIC IDEA IS TO CREATE A
CHATBOT TO RESPOND
QUESTIONS RELATED TO
TRAVELLING AND BOOKING
AND ANSWER FAQ
POSSIBLE MODELS WE CAN USE WE CAN USE IT
TOURISM WEBSITES
POSSIBLE DATA SETS FLIGHT BUS TRAIN DATA
ABOUT TIMINGS AND AVAILABILITY,HOTEL
DATA,TOURIST ATTRACTION DATA,GEOLOCATION
DATA
PRETRAINED MODEL NAMES DIALOGFLOW,
BOTPRESS, RASA, IBM WATSON ASSISTANT
PERFORMANCE METRICS RESPONSE TIME,USER
ENGAGEMENT,ACCURACY,CONVERSION
RATE,RETENTION RATE.
TECHNOLOGIES USED NLP,ML,API,CLOUD
COMPUTING,NLG(natural language
generation),CUI(conversational user interface)
TOURIST GUIDE CHATBOT ARCHITECTURE:
The chatbot as an information service can be a pillar of this concept. In
this work, we used deep learning over conventional chatbot designs
such as Pattern Matching strategy and rule-based approach. The user
input is extracted using Natural Language Processing (NLP) in a deep
learning approach. This approach doesn’t demand a predetermined
answer for every potential user query. Instead, it uses the entire
conversation context for text processing. We used the Convolutional
Neural Network (CNN), and the chatbot system receives the user’s
query, then computes the vector representations of the user’s query.
These vector representations are used as features of the neural network
to produce responses to the user.
The user can type the query in the chatbot. Then the AI chatbot engine
computes the similarity score (SS) of the user query and identifies the
suitable tag
CHATBOT ARCHITECTURE:
CHATBOT WORKFLOW:
The stakeholder community of the city includes local tourist spot
authorities, guides, local cabs, pilgrim centers, restaurants, other
small and big shops, malls, and other transportation services.
Once the user initiates a chatbot, the engine responds with a query to
understand whether the user is. If the query response falls under
tourist, the chatbot engine retrieves the latest information from the
database and responds to the user query
CHAT BOT DEEP
LEARNING MODEL:
The model has been
built using CNN,
which works as our
chatbot engine. When
the user enters a query,
the model recognizes
the language entered
by the user, gets the
corresponding user
query, and analyzes
the pattern of the
question
Based on the pattern,
it identifies the
appropriate tag of the
user query, then
searches for the
response in the databases.
Once the most suitable response is identified, it translates the
response into the user’s language and sends back the output as a
response to the user.
Once the model is built, whenever the user types a query, the model
will compute the similarity between the sequence of the new text and
the training data. The model responds to the user based on these
similarity scores (SS) for each category. The sample output we
received from the experimentation is shown
