Multilayer Neural Network Based Speech Emotion Recognition for燬mart燗ssistance

Day by day, biometric-based systems play a vital role in our daily lives. This paper proposed an intelligent assistant intended to identify emotions via voice message. A biometric system has been developed detect human based on recognition and control few electronic peripherals for alert actions. smart aims provide support the people through buzzer light emitting diodes (LED) signals it also keep track of places like households, hospitals remote areas, etc. The approach is able seven emotions: worry, surprise, neutral, sadness, happiness, hate love. key elements implementation speech emotion are processing, once recognized, machine interface automatically detects actions LED. trained tested various benchmark datasets, i.e., Ryerson Audio-Visual Database Emotional Speech Song (RAVDESS) database, Acoustic-Phonetic Continuous Corpus (TIMIT) database (Emo-DB) evaluated parameters, accuracy, error rate, time. While comparing with existing technologies, algorithm gave better rate less Error time decreased 19.79%, 5.13 s. RAVDEES dataset, 15.77%, 0.01 s Emo-DB dataset 14.88%, 3.62 TIMIT database. model shows accuracy 81.02% 84.23% 85.12% compared Gaussian Mixture Modeling(GMM) Support Vector Machine (SVM) Model.