Abstract: The disclosure discloses an epitaxial structure of a semiconductor device, as well as a manufacturing method thereof, and a semiconductor device. Therein, the epitaxial structure includes a substrate and an epitaxial layer located on one side of the substrate, and a surface roughness of one side the substrate close to the epitaxial layer is Ra, wherein 0<Ra?5 nm. In the epitaxial structure of a semiconductor device as well as the manufacturing method thereof and the semiconductor device provided by the disclosure, by setting the surface roughness Ra on one side of the epitaxial growth of the substrate to satisfy 0<Ra?5 nm, a high-quality epitaxial layer can be directly epitaxially grown on the substrate, eliminating the need to arrange a thicker buffer layer and thus reducing thermal resistance, which improves the operating performance of the semiconductor device.

Abstract: The present disclosure relates to an epitaxial structure of a semiconductor device, a device, and a method of manufacturing the epitaxial structure. The epitaxial structure includes a substrate, and a first semiconductor layer, located on the substrate, the first semiconductor layer including buffer layers, the buffer layers at least including a first buffer layer, a second buffer layer, and a third buffer layer which are arranged in layers, and the second buffer layer being located between the first buffer layer and the third buffer layer, wherein the second buffer layer is doped with iron impurities, and the first buffer layer and the third buffer layer are not actively doped with iron impurities, and concentration of iron impurities of the second buffer layer satisfies a first preset range.

Abstract: A long short-term memory (LSTM) model-based disease prediction method and apparatus, a computer device, and a storage medium are provided. The method includes: obtaining first medical data of a target object and second medical data of an associated object; inputting the first medical data and the second medical data into a first LSTM network in the LSTM model, to obtain a hidden state vector sequence in the first LSTM network; inputting the hidden state vector sequence into a second LSTM network for operation, to obtain a disease prediction result; selecting a predicted disease with an incidence rate higher than a preset threshold, and recording the predicted disease as a designated disease, and obtaining, based on a preset disease association network, an associated disease directly connected to the designated disease; and outputting the disease prediction result and the associated disease, thereby improving the prediction accuracy.

Abstract: A long short-term memory (LSTM) model-based disease prediction method and apparatus, a computer device, and a storage medium are provided. The method includes: obtaining first medical data of a target object and second medical data of an associated object; inputting the first medical data and the second medical data into a first LSTM network in the LSTM model, to obtain a hidden state vector sequence in the first LSTM network; inputting the hidden state vector sequence into a second LSTM network for operation, to obtain a disease prediction result; selecting a predicted disease with an incidence rate higher than a preset threshold, and recording the predicted disease as a designated disease, and obtaining, based on a preset disease association network, an associated disease directly connected to the designated disease; and outputting the disease prediction result and the associated disease, thereby improving the prediction accuracy.