Abstract: A system for detecting and cleaning indoor air pollution includes gas detection devices and filtering devices. The gas detection devices are adapted to detect a qualitative property and a concentration of an air pollution and output an air pollution data to perform an intelligent computation. The filtering devices are physical-typed or chemical-typed for filtering the air pollution. The filtering devices include one or more movable filtering devices, and the movable filtering device includes a gas detection device. After the intelligent computation is performed to locate an air pollution location, a control command is transmitted to the movable filtering device selectively and intelligently, and the movable filtering device receives the control command and is moved to the air pollution location. Therefore, the movable filtering device allows the air pollution data to approach to a non-detection state, thus a gas in the indoor space is cleaned to a safe and breathable state.

Abstract: The present invention provides a method and device for comparing similarities of high dimensional features of images, capable of improving the retrieval speed and retrieval precision in a similarity retrieval from massive images based on Locality Sensitive HASH (LSH) code. The method for comparing similarities of high dimensional features of images according to the present invention comprises: reducing dimensions of extracted eigenvectors of the images by the LSH algorithm to obtain low dimensional eigenvectors; averagely segmenting the low dimensional eigenvectors and establishing a segment index table; retrieving the segmented low dimensional eigenvector of a queried image from the segment index table to obtain a candidate sample set; and performing a similarity metric between a sample in the candidate sample set and the low dimensional eigenvector of the queried image.

Abstract: An Embodiment of the present disclosure provides a personalized search device based on product image features, comprising a feature extraction module configured to extract, using a neural network model, an abstract semantic feature vector of an image by category, a category image calculation module configured to calculate a mean and a variance of the abstract semantic feature vector respectively for each dimension, and perform normalization processing, in each dimension, on the abstract semantic feature vector; a user browsing behavior weight calculation module configured to sum the normalized abstract semantic feature vectors extracted by category from all the images browsed by a user, so as to obtain an interest weighting vector of the user for each category; a ranking module configured to get, according to the interest weighting vector of each user for a category, an inner product on feature vectors of images not viewed by the user for the category, so as to obtain a score of each of the images; rank the i

Abstract: The present invention provides a method and device for comparing similarities of high dimensional features of images, capable of improving the retrieval speed and retrieval precision in a similarity retrieval from massive images based on Locality Sensitive HASH (LSH) code. The method for comparing similarities of high dimensional features of images according to the present invention comprises: reducing dimensions of extracted eigenvectors of the images by the LSH algorithm to obtain low dimensional eigenvectors; averagely segmenting the low dimensional eigenvectors and establishing a segment index table; retrieving the segmented low dimensional eigenvector of a queried image from the segment index table to obtain a candidate sample set; and performing a similarity metric between a sample in the candidate sample set and the low dimensional eigenvector of the queried image.

Abstract: A method and device for determining whether a user who has not ordered a commodity has a demand for the commodity. The method comprises calculating a number of actions directed at the commodity by users in a preselected time period that is not ordered in a preselected time period and a number of users purchasing the commodity after the preselected time period; establishing a training set based on the numbers and a model corresponding to the training set. The model has an input value of the number of actions directed to the commodity by a user and an output value of whether the user purchases the specified commodity. The method also includes calculating the number of actions of an object user who has not ordered in a preset time period and inputting the number into the model as the input value to obtain the output value of the model.

Abstract: Method for judging age brackets of users including acquiring consumption data of users and establishing models based on the consumption data. Establishing the models includes dividing the consumption data into training data and test data, calculating a number of the users of the training data in predetermined age brackets, calculating a number of each tertiary category of the training data in the predetermined age brackets, calculating probabilities that each tuple of the test data belongs to each of the predetermined age brackets based on the number of the users and the number of the tertiary categories, selecting the age bracket with the maximum probability as the age bracket to which the user corresponding to the tuple belongs, comparing errors between the predetermined age brackets and the selected age bracket to obtain a predictive error rate, and outputting models with predictive error rates larger than or equal to a predetermined threshold.