An automated respiration and heart rate estimation system through signal change analysis would be helpful for early disease prediction. Traditional methods frequently require attached sensors, rendering them unsuitable for continuous monitoring in home environments. The existing channel state (CSI) information based works did not investigate both heart and respiration rate prediction using both machine (ML) and deep learning (DL) techniques with higher accuracy. To outperform these issues, this paper takes advantage of the ubiquitous nature of Wi-Fi signals to estimate human respiration (RR) and heart rate (HR) using Wi-Fi CSI data. The proposed system captures CSI data using standard Wi-Fi hardware, demonstrating how the human body influences electromagnetic signals as they travel through space. A new feature extraction technique that combines shapelet transform and Fast Fourier Transform (FFT) is introduced to improve signal representation and allow for accurate estimation of vital signs. Several ML and DL learning models were implemented and rigorously evaluated. The simulation results hinted that the Extra Trees model shows higher R-squared and lower MAE than other ML and DL models. The comparison results showed that the proposed scheme has at least 5% higher R-Squared and 24% lower MAE value than previous works