The invention provides a
urination early warning method and
system for
time sequence prediction based on a large
language model, and the method at least comprises the following steps: S1, collecting
bladder volume dynamic change data through a
urine volume collection terminal in a non-contact manner, and transmitting the data to a mobile terminal APP through a
Bluetooth transmission unit; s2, preprocessing the original volume data acquired by the
urine volume acquisition terminal through the mobile terminal APP, and recording static characteristics of a user; s3, carrying out
feature fusion on the preprocessed volume
time sequence data and static features through a large
language model prediction
server, constructing a three-dimensional input matrix, carrying out prediction reasoning by utilizing a pre-trained large
language model, and outputting a future
urination time point; and S4, triggering early warning and reminding according to the prediction result and the real-time
bladder capacity. The
system at least comprises a
urine volume collection terminal, a mobile terminal APP and a large language
model prediction server. According to the method and the
system, accurate prediction of the
urination time in the future is realized through
deep learning of individual historical data by a large language model.