Method for denoising click data, electronic device and storage medium
A method for denoising click data includes: acquiring a set of click data including pieces of first click data and a real label corresponding to each piece of first click data; extracting feature vectors of each piece of first click data with a graph model; dividing the feature vectors into sets of feature vectors; obtaining trained binary classification models by training binary classification models with the sets of feature vectors; for each of the feature vectors, obtaining prediction values corresponding to the feature vector by predicting the feature vector with the trained binary classification models, and calculating a prediction label of the feature vector based on the prediction values of the feature vector; and removing noise data in the pieces of first click data, based on the pieces of first click data, the real label and the prediction label of each piece of first click data.
1. A method for denoising click data in big data and cloud services, performed by an electronic device, comprising:
acquiring a user identifier and a recommended content identifier associated with each piece of first click data, wherein each piece of the first click data indicates a piece of content is recommended to a user;
acquiring user data and content data associated with each piece of the first click data based on the user identifier and the recommended content identifier;
extracting a feature vector of each piece of the first click data with a graph model;
dividing a plurality of feature vectors into sets of feature vectors;
obtaining trained binary classification models by training binary classification models with the sets of feature vectors;
for each of the plurality of feature vectors, obtaining prediction values of the feature vector by predicting the feature vector with the trained binary classification models;
for each of the plurality of feature vectors, calculating a prediction label of the feature vector based on the prediction values of the feature vector;
in response to determining that a real label corresponding to each piece of the first click data indicates that the user clicks the piece of content recommended, determining the piece of first click data as a piece of second click data; and
removing noise data in pieces of second click data based on prediction labels of the pieces of second click data, wherein the noise data refers to that the user casually clicks the piece of content recommended.
2. The method of claim 1 , wherein before extracting the feature vector of each piece of the first click data with the graph model, the method further comprises:
acquiring a training set of click data; and
training the graph model with the training set of click data.
3. The method of claim 1 , wherein training the binary classification models with the sets of feature vectors, comprises:
for each time, training a different binary classification model from the binary classification models, with a set of feature vectors from the sets of feature vectors as a verification data set and remaining sets of feature vectors from the sets of feature vectors as training data sets, wherein the verification data set is different at each time.
4. An electronic device, comprising:
at least one processor; and
a memory, stored with instructions executable by the at least one processor, wherein the at least one processor is caused to:
acquiring a user identifier and a recommended content identifier associated with each piece of first click data, wherein each piece of the first click data indicates a piece of content is recommended to a user;
acquiring user data and content data associated with each piece of the first click data based on the user identifier and the recommended content identifier;
extracting a feature vector of each piece of the first click data with a graph model;
dividing a plurality of feature vectors into sets of feature vectors;
obtaining trained binary classification models by training binary classification models with the sets of feature vectors;
for each of the plurality of feature vectors, obtaining prediction values of the feature vector by predicting the feature vector with the trained binary classification models;
for each of the plurality of feature vectors, calculating a prediction label of the feature vector based on the prediction values of the feature vector;
in response to determining that a real label corresponding to each piece of the first click data indicates that the user clicks the piece of content recommended, determining the piece of first click data as a piece of second click data; and
removing noise data in pieces of second click data based on prediction labels of the pieces of second click data, wherein the noise data refers to that the user casually clicks the piece of content recommended removing.
5. The electronic device of claim 4 , wherein before extracting the feature vector of each piece of the first click data with the graph model, the at least one processor is caused to:
acquiring a training set of click data; and
training the graph model with the training set of click data.
6. The electronic device of claim 4 , wherein training the binary classification models with the sets of feature vectors, comprises:
for each time, training a different binary classification model from the binary classification models, with a set of feature vectors from the sets of feature vectors as a verification data set and remaining sets of feature vectors from the sets of feature vectors as training data sets, wherein the verification data set is different at each time.
7. A non-transitory computer-readable storage medium stored with computer instructions, wherein, the computer instructions are configured to cause a computer to perform a method for denoising click data, the method comprising:
acquiring a user identifier and a recommended content identifier associated with each piece of first click data, wherein each piece of the first click data indicates a piece of content is recommended to a user;
acquiring user data and content data associated with each piece of the first click data based on the user identifier and the recommended content identifier;
extracting a feature vector of each piece of the first click data with a graph model;
dividing a plurality of feature vectors into sets of feature vectors;
obtaining trained binary classification models by training binary classification models with the sets of feature vectors;
for each of the plurality of feature vectors, obtaining prediction values of the feature vector by predicting the feature vector with the trained binary classification models;
for each of the plurality of feature vectors, calculating a prediction label of the feature vector based on the prediction values of the feature vector;
in response to determining that a real label corresponding to each piece of the first click data indicates that the user clicks the piece of content recommended, determining the piece of first click data as a piece of second click data; and
removing noise data in pieces of second click data based on prediction labels of the pieces of second click data, wherein the noise data refers to that the user casually clicks the piece of content recommended.
8. The storage medium of claim 7 , wherein before extracting the feature vector of each piece of the first click data with the graph model, the method further comprises:
acquiring a training set of click data; and
training the graph model with the training set of click data.
9. The storage medium of claim 7 , wherein training the binary classification models with the sets of feature vectors, comprises:
for each time, training a different binary classification model from the binary classification models, with a set of feature vectors from the sets of feature vectors as a verification data set and remaining sets of feature vectors from the sets of feature vectors as training data sets, wherein the verification data set is different at each time.