Aim: The purpose of this study's work is to improve the accuracy of predicting depression

using evaluating Convolutional Neural Networks and Vanilla Neural Networks. The
examination compares the effectiveness of Convolutional Neural Network (CNN) and Vanilla
Neural Network in improving the chatbot's capacity to understand and reply to user’s
emotional states. Materials and Methods: In these studies, the contrast of organizations is taken
into consideration, Convolutional Neural Networks (N=five) and Vanilla Recurrent Neural
Networks (N=5) are executed with various education at the preprocessed dataset. Materials and
Methods: In this research comparison of two groups was taken into consideration,
Convolutional Neural Networks (N=5) and Vanilla Recurrent Neural Networks (N=5) are
executed with varying training on the preprocessed dataset. The models were evaluated using
various metrics, including accuracy, and loss, to assess their performance in detecting signs of
depression in text data for predicting depression with the efficient method. The G-power test
used is about 85% (g power setting parameters: α=0.05 and power=0.85). Result: The results of
this study demonstrate that the CNN model is the superior choice for depression detection in
text data, achieving a remarkable accuracy rate of 95.28% compared to the VRNN model's
46.47%. These findings emphasize the critical role of model selection in the success of natural
language processing projects, offering valuable insights for future applications of AI in mental
health assessment and support. Conclusion: In this study, the comparative analysis between
Convolutional Neural Networks and Vanilla Recurrent Neural Networks for Depression
Detection has led to the conclusion that Convolutional Neural Networks exhibit superior
accuracy. Keywords: Convolution neural networks, Recurrent neural networks, Depression, Mental
health, Natural language, Chatbot.

INTRODUCTION
In an increasing number of digitized global, the synergy of technology and intellectual health is
a subject of paramount importance. Depression, a pervasive and debilitating intellectual fitness
situation affecting tens of millions globally, regularly stays undetected or untreated. This task,
entitled "A Smart AI Companion," tackles the urgent difficulty of identifying and addressing
depression (Fadul et al. 2023) (Xu et al. 2023) through innovative chatbot solutions, offering a
comparative evaluation among Convolutional Neural Networks (CNN) and Vanilla Neural
Networks. In an age characterized by the pervasive effect on generation, the ability of artificial
intelligence (AI) in intellectual health care cannot be underestimated. Depression is a
complicated sickness with a wide spectrum of signs and symptoms, rendering it a hard
condition to diagnose. Early detection and intervention are essential to ameliorating its effects
and improving the lifestyles of affected individuals. The creation of chatbots and Natural
Language Processing (NLP) techniques (Li et al. 2023) has opened up new horizons for
intellectual fitness help, providing a discreet, handy, and non-judgmental platform for
individuals to articulate their emotions (Hentrich 2021).
These AI-driven chatbots possess the capability to engage in conversations, scrutinize textual
inputs, and monitor emotional well-being over time. This project aspires to empower
individuals with an intelligent AI companion proficient in detecting signs of depression,
rendering support, and, when necessary, facilitating connections to mental health(Haug and
Kurpicz-Briki 2022) professionals for further assistance. The research undertaken in this
project delves into a comparative evaluation of two neural network architectures,
Convolutional Neural Networks (CNN) and Vanilla Neural Networks, in the analysis of text-
based data(S. Kim et al. 2023) for depression detection(Milintsevich, Sirts, and Dias 2023). This
comparative analysis furnishes valuable insights into the potential advantages of leveraging
CNN, renowned for its prowess in image processing, for the analysis of textual content in the
context of mental health assessments(Wright-Berryman et al. 2023). The objective is to enhance
the accuracy and efficiency of depression detection, ensuring that individuals receive timely
and suitable support. Situated at the intersection of AI and intellectual fitness, this venture
holds the promise of a greater connected, empathetic, and technologically advanced approach
to addressing despair within the virtual era.

The proliferation of studies in the discipline of AI-powered mental health assist is evident in the
body of work emphasizing the efficacy of peer support interventions for depression. Moreover,
the growing reliance on social media(Turner et al. 2022)as a platform for individuals to express
their thoughts and emotions has led to the development of algorithms designed to quantify
mental health signals. Deep learning techniques(A. Y. Kim et al. 2023), represented by neural
networks, have been a driving force behind AI advancements and have found application in
diverse fields, including mental health analysis. These references collectively underscore the
significance of the "A Smart AI Companion" project in the context of contemporary mental
health research and technology development. This study investigates the efficacy of
Convolutional Neural Network (CNN) and Vanilla Neural Network fashions, aiming to enhance
the chatbot's empathetic capabilities in expertise and responding to users' emotional proper
being. The research delves into the comparative analysis of these neural network architectures
to optimize mental health support(Francese and Attanasio 2023) through artificial intelligence.

MATERIALS AND METHODS

The research carried out at Saveetha School of Engineering's Machine Learning Lab in
Chennai centered on developing a revolutionary Smart AI companion able to detect and
address depression through chatbot interactions. The primary objective was to compare the
effectiveness of Convolutional Neural Networks (CNN) and Vanilla Neural Networks (VNN) in
enhancing the chatbot's ability to understand and respond to user's emotional states. The study
employed Gpower software to determine a sample size of 85% accuracy for a robust comparison
between CNN and VNN. Two algorithms were implemented using technical analysis software to
assess their impact on the chatbot's performance in detecting and addressing depression. The
methodology prioritized ethical considerations, and ethical approval was deemed unnecessary,
indicating adherence to ethical standards and minimal risk to participants. The research underscores a
commitment to leveraging advanced technology and scientific rigor to enhance the capabilities of AI
in addressing mental health challenges. This study contributes valuable insights into optimizing
neural network algorithms for empathetic and effective user interactions, marking a significant
advancement in the intersection of artificial intelligence and mental health support.

The proposed work is designed and implemented with the help of Python OpenCV software. The
platform to assess deep learning was Windows 10 OS. The hardware configuration was an Intel core
i7 processor with a RAM size of 4GB. The system sort used was 64-bit. For the implementation of
code, the Python programming language was used. As for code execution, the dataset is worked
behind to perform an output process for accuracy. For the research project on detecting and
addressing depression through chatbot insights, a dataset was sourced from Kaggle, a prominent
platform for datasets and machine learning resources. The dataset, structured in JSON format, was
curated to facilitate the comparison of Convolutional Neural Network (CNN) and Vanilla Neural
Network (VNN) models. This dataset contains a diverse array of user interactions, capturing textual
and emotional nuances relevant to depression detection. Attributes include text messages, sentiment
labels, and contextual information crucial for training and evaluating the chatbot's performance(Haug
and Kurpicz-Briki 2022). To execute the machine learning models seamlessly, the research was
conducted in Google Colab, a cloud-based platform that allows for collaborative coding and
execution of Python scripts. This integration with Google Colab not only facilitated efficient
processing of the dataset but also streamlined the implementation and comparison of the CNN and
VRNN(Milintsevich, Sirts, and Dias 2023) models, contributing to the research's computational
efficiency and reproducibility.

Convolutional Neural Networks Algorithm:

Description:

The Convolutional Neural Network (CNN)(Bertl et al. 2024) is a deep learning architecture primarily
designed for image processing tasks. In the context of text analysis, it can be adapted for feature
extraction by treating the text as an image, with one dimension representing word position and the
other representing word embeddings or vectors. The CNN model for text data consists of multiple
layers. The Embedding Layer converts words into dense vectors, which serve as the input for the
CNN model. Convolutional layers use a set of learnable filters to convolve over the embedded words,
capturing local patterns and features. This is particularly effective for identifying n-grams (sequences
of n words) in the text. Max-pooling or average-pooling layers follow the convolutional layers,
reducing the dimensionality of the extracted features while retaining important information. After
pooling, fully connected layers are used for classification, making predictions based on the learned
features.

Algorithm:

1. Input JSON dataset for a chatbot with depression labels.

2. Let ‘f’ be the feature 3D matrix.

3. Vectorize the data in the ‘f’ dataset.

4. Split the feature set and labels.

5. Apply Convolutional Neural Networks algorithm.

6. Train the model using the training set.

7. Evaluate the model using the training set.

8. Return the accuracy and loss.

Vanilla Recurrent Neural Network (RNN):

The Vanilla Recurrent Neural Network (RNN)(Jose et al. 2020) is a fundamental deep learning
architecture for sequential data, making it well-suited for natural language processing tasks. Unlike
CNN, RNNs capture dependencies over time, making them suitable for analyzing the sequential
nature of text data. The Vanilla RNN model consists of various components. The embedding layer
converts words into continuous vectors. RNN Layers contain recurrent units, typically implemented
as Long Short-Term Memory (LSTM) or Gated Recurrent Unit (GRU) cells. These units enable the
network to capture sequential dependencies by maintaining hidden states that evolve as new words
are processed. A fully connected layer at the output processes the final hidden state and makes
predictions. In the context of this project, the output represents the probability of the input text
indicating signs of depression.

Algorithm:

1. Input JSON dataset for a chatbot with depression labels.

2. Let ‘f’ be the feature 3D matrix.

3. Vectorize the data in the ‘f’ dataset.

4. Split the feature set and labels.

5. Apply the Vanilla Recurrent Neural Networks algorithm.

6. Train the model using the training set.

7. Evaluate the model using the training set.

8. Return the accuracy and loss.

Statistical Analysis

Statistical analysis of Novel Convolutional Neural Networks and the Inception Algorithm involves
using SPSS software(Alemayehu et al. 2023). Data serves as the independent variable, while
accuracy is the dependent variable. The analysis includes conducting an Independent T-test to
calculate the accuracy of both methods. Statistical analysis is a vital aspect of the " A Smart AI
companion: detecting and addressing depression with chatbot insights by comparing CNN to Vanilla
Neural Network " project, as it serves to evaluate the performance and effectiveness of the employed
machine learning models—Convolutional Neural Network (CNN) and Vanilla Neural Network. In
this section, we will discuss the statistical methods and key metrics used for model evaluation, as
well as the implications of the results.

RESULTS

The comparative analysis of the Convolutional Neural Network (CNN) and Vanilla Recurrent Neural
Network (VRNN) models for depression detection yielded striking differences in performance. The
CNN model exhibited a notably high accuracy rate of 95.28%, while the VRNN model, although still
promising, achieved a considerably lower accuracy rate of 46.47%. These findings underline the
significant impact of model architecture on the ability to detect signs of depression in text data.

Table 1 presents a summary of the performance metrics for both the CNN and VRNN models. The
stark contrast in accuracy rates is apparent, with the CNN model far outperforming the VRNN model.
While the CNN model demonstrates a balanced performance across accuracy and loss, the VRNN
model's metrics are comparatively lower, indicating a struggle to correctly classify depression-related
text content. The bar chart clearly illustrates the substantial difference in accuracy, with the CNN
model significantly outshining the VRNN model. This visualization underscores the importance of
selecting an appropriate deep learning architecture when undertaking natural language processing
tasks such as depression detection.

DISCUSSION

The results of this study underscore the critical importance of selecting the appropriate deep learning
architecture for natural language processing tasks, particularly in the context of depression detection.
The discussion below delves into the implications and potential explanations for the notable
performance difference between the Convolutional Neural Network (CNN)(Ksibi et al. 2023) and the
Vanilla Recurrent Neural Network (VRNN) models. The stark contrast in accuracy rates between the
CNN(Tigga and Garg 2023) and VRNN models is primarily attributed to their architectural
differences. CNNs, designed for image processing tasks, are known for their effectiveness in
capturing local patterns and features, even when applied to textual data. In contrast, VRNNs, while
proficient at handling sequential data, may struggle with the complex dependencies and patterns
inherent in natural language, such as those indicative of depression. The highly sequential and
contextual nature of text data makes it a challenging domain for VRNNs, leading to their lower
accuracy in this study.

One of the strengths of the CNN(Wang et al. 2023) model is its ability to efficiently extract features
from text data by treating it as a spatial image. The application of convolutional layers allows the
model to identify significant textual patterns and representations, which are crucial for detecting
signs of depression. The VRNN, in contrast, may not effectively capture these patterns due to its
limited ability to handle the spatial aspects of text. Another contributing factor to CNN's superior
performance could be the size of the training dataset. Larger datasets often provide a richer variety of
examples and patterns for the model to learn from. In cases where the VRNN was less accurate, it
may have been due to a reduced ability to generalize effectively from the available data, whereas
CNN's feature extraction capabilities allowed it to excel even with a limited dataset.

Depression detection, through the analysis of textual content, involves understanding not only
individual words but also the context and nuances of language. CNN(Nadeem et al. 2022) models can
excel at recognizing subtle textual cues and emotional nuances, while VRNN models may struggle to
maintain context over longer sequences of text, potentially leading to a loss of important information.
The findings of this study have substantial implications for the design of AI-driven systems for
mental health support. The selection of an appropriate model architecture can significantly impact the
accuracy and efficacy of such systems. Future research may explore hybrid models or advanced
architectures that combine the strengths of both CNN and VRNN, to achieve even higher accuracy
and a more nuanced understanding of text data.

CONCLUSION

In summary, the results of this study demonstrate that the CNN model is the superior choice for
depression detection in text data, achieving a remarkable accuracy rate of 96% compared to the
VRNN model's 48%. These findings emphasize the critical role of model selection in the success of
natural language processing projects, offering valuable insights for future applications of AI in
mental health assessment and support.