This disclosure describes one or more implementations of a model segmentation system that generates accurate audience segments for client devices/individuals utilizing multi-class decision tree machine-learning models. For example, in various implementations, the model segmentation system generates a customized loss penalty matrix from multiple loss penalty matrices. In particular, the model segmentation system can generate regression mappings of model evaluation metrics for a plurality of decision tree models and combine loss penalty matrices based on the regression mappings to generate a customized loss penalty matrix that best fits an administrator's customized needs of segment accuracy and reach. The model segmentation system then utilizes the customized loss penalty matrix to train a multi-class decision tree machine-learning model to classify client devices into non-overlapping audience segments. Further, in one or more implementations, the model segmentation system refines the multi-class decision tree machine-learning model based on adjusting the tree depth.

Systems, methods, and computer-readable media are disclosed for filtering data with probabilistic filters for content selection. In one embodiment, an example method may include determining a user interaction history with a first product identifier for a user account, determining a first parent product identifier of the first product identifier, and generating a database with the first parent product identifier and a user account identifier for the user account. Example methods may include determining a set of candidate content with first content and second content for the user account, determining a second product identifier associated with the first content, and determining a second parent product identifier of the second product identifier. Example methods may include determining that the second parent product identifier is not present in the database using a probabilistic filter, and determining that the first content is eligible for presentation.

A system and method is provided for delivering customized information related to a specific product of interest to consumer. In practice, the consumer uses a suitably enabled portable, mobile and/or wireless device (e.g., such as a mobile camera phone) to scan or otherwise read a marker associated with a product, retail item or other article of interest. From the marker, a unique ID is obtained (i.e., the marker ID). The marker ID is then used to cross-reference a URL or other like address in a database that relates marker IDs to corresponding URLs. The target URL is returned to the consumer's device and an http session is established with a content management server at the target URL. In one suitable embodiment, the content management server obtains a SKU and/or template web page ID that are associated with the marker ID in a database. Suitably, the SKU relates to the specific product with which the marker was associated. Having in this way determined the actual specific product of interest to the consumer, customized information related thereto can be delivered to the consumer's device, e.g., via a web page optimized for the identified device. Optionally, the web page content and/or template is obtained from a database that associates the same with the template web page ID received by the content management server.

Systems and methods of generating, for display on a graphical user interface (GUI), a personalized advertisement. The systems and methods can include receiving data indicative of initiating a browsing session, aggregating user activity data associated with a user of the application, the user activity data including current session data and/or past session data, applying machine learning on the user activity data to generate one or more excluded products, applying the one or more excluded products to a product database to generate a list of relevant products, ranking, using one or more ranking rules, the list of relevant products to generate a ranked list of relevant products, and sending, to the mobile device, the personalized advertisement including one or more relevant products from the ranked list of relevant products.

The present disclosure is directed to facilitating privacy conscious market collaboration. A first-party computing system can access a first-party user information attribute from a first-party identified user profile associated with a first-party identified user and generate a first-party hashed user information attribute including indecipherable text by applying a predetermined hash function. The system can transmit to a third-party computing system a communication including the first-party hashed user information attribute and a payload including customer insight data. The third-party computing system can determine that a third-party hashed user information attribute associated with a third-party identified user profile includes indecipherable text that matches the indecipherable text of the first-party hashed user information attribute. The third-party computing system can provide to a third-party identified user associated with the third-party identified user profile an advertisement that is based on the customer insight data in the payload of the communication.

A computer system for dynamically generating and displaying a product recommendation to a user by analyzing multidimensional attributes associated with users and physical items to generate a unique score for the physical items is described. In one example, a computer system includes an attribute analysis engine to analyze user profile data and assign at least one multidimensional attribute to the user. The attribute analysis engine may also analyze physical items data and assign at least one multidimensional attribute to the physical items. The multidimensional attribute assigned to the physical items may match the multidimensional attribute assigned to the user. In one example, the computer system can dynamically generate a score for the physical items based on the comparison of the multidimensional attribute assigned to the physical items and the multidimensional attribute assigned to the user.

The present disclosure is directed to facilitating privacy conscious market collaboration. A first-party computing system can access a first-party user information attribute from a first-party identified user profile associated with a first-party identified user and generate a first-party hashed user information attribute including indecipherable text by applying a predetermined hash function. The system can transmit to a third-party computing system a communication including the first-party hashed user information attribute and a payload including customer insight data. The third-party computing system can determine that a third-party hashed user information attribute associated with a third-party identified user profile includes indecipherable text that matches the indecipherable text of the first-party hashed user information attribute. The third-party computing system can provide to a third-party identified user associated with the third-party identified user profile an advertisement that is based on the customer insight data in the payload of the communication.

Systems and methods for personalization using image clouds to represent content. Image clouds can be used to identify initial user interest, present recommended content, present popular content, present search results, and present user profile information. Image clouds are interactive, allowing users to select images displayed in the image cloud, which can contribute to presenting more personalized content as well as updating a user's profile.

A computer-implemented method for identifying directed content without access to personally-identifiable information of a user includes receiving a group identifier that identifies a group to which the user belongs and an identifier for a device of the user; selecting content that is determined to be responsive to preferences of the group, without using information that identifies the user; and providing the selected content for display on the device of the user.

In a method for delivering targeted television advertisements based on online behavior, IP addresses indicating online access devices and IP addresses indicating television set-top boxes are electronically associated for a multitude of users. Using user profile information derived from online activity from one of the online access IP addresses, a television advertisement is selected, such as by using behavioral targeting or demographic information, and automatically directed to the set-top box indicated by the set-top IP address associated with that online access IP address. Preferably neither the user profile information nor the electronic association of online access and set-top box IP addresses includes personally identifiable information.