International Journal of Trend in Scientific Research and Development

As everyone knows that Sentimental analysis plays an important role in these days because many start ups have started with user driven content 1 . Only finding the voice is not be the real time scenario so finding the Sentiment analysis of agent and customer separately is an important research area in natural language processing. Natural language processing has a wide range of applications like voice recognition, machine translation, product review, aspect oriented product analysis, sentiment analysis and text classification etc 2 . This process will improve the business by analyze the emotions of the conversation with respect to the customer voice separately and also agent voice separately. In this project author going to perform speaker identification and analyze the sentiment of the customer and agent separately using Amazon Comprehend. Amazon Comprehend is a natural language processing NLP service that uses machine learning to extract the content of the voice. By using the speaker identification author can extract the unstructured data like images, voice etc separately so it is easy to analyze the business performance. Thus, will identify the emotions of the conversation and give the output whether the customer conversation is Positive, Negative, Neutral, or Mixed. To perform this author going to use some services from Aws due to some advantages like scaling the resources is easy compare to the normal process like doing physically such as support vector machine SVM . AWS services like s3 is a object data store, Transcribe which generate the audio to text in raw format, Aws Glue is a ETL Service which will extract transform and load the data from the S3, Aws Comprehend is a NLP service used for finding sentiment of audio, Lambda is a server less where author can write a code, Aws Athena is a analyzing tools which will make complex queries in less time and last there is quick sight is a business intelligent tool where author can visualize the data of customers and also agents. 

This paper is aimed to implement Classification of Language Speech Recognition System by using feature extraction and classification. It is an Automatic language Speech Recognition system. This system is a software architecture which outputs digits from the input speech signals. The system is emphasized on Speaker Dependent Isolated Word Recognition System. To implement this system, a good quality microphone is required to record the speech signals. This system contains two main modules feature extraction and feature matching. Feature extraction is the process of extracting a small amount of data from the voice signal that can later be used to represent each speech signal. Feature matching involves the actual procedure to identify the unknown speech signal by comparing extracted features from the voice input of a set of known speech signals and the decision making process. In this system, the Mel frequency Cepstrum Coefficient MFCC is used for feature extraction and Vector Quantization VQ which uses the LBG algorithm is used for feature matching.