Multi-dimensional interest fusion recommendation method and device based on neural network and related equipment
The invention discloses a multi-dimensional interest fusion recommendation method and device based on a neural network and related equipment, and the method comprises the steps: inputting click data into an interest fusion model which comprises a main interest model, an auxiliary interest model and...
Gespeichert in:
Hauptverfasser: | , , |
---|---|
Format: | Patent |
Sprache: | chi ; eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | The invention discloses a multi-dimensional interest fusion recommendation method and device based on a neural network and related equipment, and the method comprises the steps: inputting click data into an interest fusion model which comprises a main interest model, an auxiliary interest model and an interest interference model, the main interest model is a graph neural network model, and the auxiliary interest model is a graph neural network model; the affiliated interest model is a recurrent neural network model, and the interest interference model is a deep belief network; fusing the main interest model and the subsidiary interest model through a multi-layer perceptron to obtain a multi-dimensional hybrid model, and training the click data by adopting the multi-dimensional hybrid model to obtain a first interest sequence; verifying the first interest sequence by adopting an interest interference model and the click data, and determining a second interest sequence according to a verification result; and pe |
---|