Travel demand predicting models are one of the widely used models that are used by researchers all over the world, and in these models, people's socio-economic characteristics are used as input to the model, and these characteristics were collected through various statistical methods. With the growth of the passenger population, population-based big data is widely used in transportation planning, and traditional methods are not effective for collecting socio-economic characteristics of people. Therefore, the researchers decided to obtain these parameters by using the pattern of trips that people have already made. In this study, with the help of artificial intelligence models and especially deep learning models, it has been tried to extract these features from the travel patterns of people. The LSTM model is one of the powerful models used to remember past events. Age, income and gender are the most important inputs of travel demand predicting models. Different age, gender and income groups may have specific travel behavior preferences, which are tried to be discovered by examining their travel behaviors. One of the important sources for determining people's travel patterns is mobile phone data that shows their daily movements. Due to the unavailability of travelers' mobile phone data, the NHTS National Household Travel Survey survey data set was used in this research. Using the obtained results, with the help of mobile phone data of passengers, their socio-economic characteristics are obtained for use in travel demand models.