Sentiment classification in Big Data involves the utilization of natural language processing (NLP) and machine learning (ML) models to analyze extensive textual data and determine the expressed sentiment. By utilizing distributed computing frameworks such as Apache Hadoop or Spark, sentiment analysis models can efficiently process large datasets, identifying sentiments like negative, positive, or neutral. This approach enables organizations to extract valuable insights from vast amounts of user-generated content, social media posts, and consumer reviews, contributing to informed decision-making, brand management, and user satisfaction analysis on a scale that traditional models would find challenging to handle. This manuscript offers the design of Chaotic Harris Hawks Optimization with Deep Autoencoder-based Sentiment Classification (CHHODAE-SC) technique on the big data environment. The main objective of CHHODAE-SC method is to detect various types of sentiments that exist in the big data environment. At a nearly stage, data preprocessing is applied to convert the input dataset into useful format. In addition, the Glove approach is employed for word embedding purposes. For the classification of sentiments, a deep autoencoder (DAE) model can be applied. At last, the CHHO algorithm is used for the optimum hyper parameter selection of DAE algorithm. Furthermore, Apache Spark is used for handling big data. Sequences of experimentations were involved to exhibit the improved detection outcomes of the CHHODAE-SC method. The experimental values highlighted that the CHHODAE-SC technique gains improved performance over existing approaches.