| CPC F24F 11/64 (2018.01) [F24F 11/526 (2018.01); G05B 13/027 (2013.01); G06F 40/279 (2020.01); G06V 30/422 (2022.01)] | 7 Claims |
|
1. A method for controlling an air conditioner, comprising:
receiving, by a reception device, data collected by a control device during an operation of an air conditioner, wherein the data comprises question information, and at least one of indoor image data and an operation status diagram of the air conditioner located indoors, the question information comprises: questions put forward by operators is collected during the operation of the air conditioner, and the question information is proposed when a user controls the air conditioner according to a voice control instruction;
inputting, by a processor, the collected data to a training model to obtain an answer that corresponds to the question information; and
playing, by a control device, the answer;
wherein before the inputting the collected data to the training model to obtain the answer that corresponds to the question information, the method further comprises: obtaining a training data set, wherein the training data set comprises at least one of the operation status diagram of the air conditioner, a collected opening and closing status diagram of indoor doors and windows and question and answer information; and generating the training model on the basis of the training data set;
wherein the generating the training model on the basis of the training data set comprises: pre-processing the training data set to obtain a pre-processing result, and processing the pre-processing result to obtain the training model, which comprise: extracting image features of the operation status diagram and/or the opening and closing status diagram by using a convolutional neural network, converting the question and answer information into a word vector list by using a word vector technology, and extracting question features from the word vector list; performing element dot product processing on the image features and the question features; performing classification operation on a result after element dot product processing to obtain an answer prediction value; and storing the image features, the question features and the answer prediction value associated with the image features and the question features to generate the training model, wherein before the playing the answer, the method further comprises:
converting the answer into voice information by the control device; and
playing the voice information by the control device and receiving adjustment information input by a target object on the basis of the voice information, wherein the adjustment information is used to adjust an operation status of the air conditioner.
|