Multimodal Sentiment Analysis for social networks with Risk Factor for Detecting Maternal Health Issues using ML and DL Classifications - A Comparative Study

Abstract

Postnatal care includes the care given to a woman after giving birth, whereas prenatal care is the care provided to a mother throughout pregnancy. Routine checks, testing, and medical guidance are all part of prenatal care. Exams, immunizations, and assistance for the mother and child in the initial weeks following birth are all included in postnatal care. Frequent check-ups can lower the chance of consequences by identifying any possible issues early on. Pregnancy could lead to potential issues, so it's essential to be informed of the risks and implications that could arise. Sentiment analysis (SA) has gained much attraction in the field of artificial intelligence (AI) and natural language processing (NLP). MSA utilizes latest advancements in machine learning and deep learning at various stages including for multimodal feature extraction and fusion and sentiment polarity detection, with aims to minimize error rate and improve performance. Latest advancements in machine learning and deep learning with various stages of processing tends to improve the performance and minimize error rate. Different ML and DL techniques classifies risk level of prenatal and postnatal pregnancy. In this paper, it addresses the problem of sentiment classification on the maternal health risk data to perform deep learning and machine learning approach. Prenatal and postnatal pregnancy state indicates whether the patient is in normal or abnormal condition. Real-time modalities of real-time maternal risk are exploited to analyze the structure of emotions implied by multimodal analysis. Different techniques such as LR, RF, SVM, Naïve Bayes, Decision Tree are analyzed in ML and in DL use Keras with TensorFlow backend XGBoost, CNN, LSTM and multi-layer perception with multimodal sentiment analysis are implemented. According to the outcomes, MLP (Multi-layer Perceptron) with multimodal sentiment analysis (MSA) enhances the maternal risk sentiment analysis's 89% accuracy and overall precision.