16 results about "Retention rate" patented technology

The invention relates to the technical field of computers, and discloses a member strategy processing method, system and device and a storage medium. The method comprises the following steps: collecting target information of a user, wherein the target information comprises historical payment behavior data, historical use behavior data and fund change data of a platform account; based on the historical payment behavior data and fund change data of a platform account, obtaining a payment capability level of the user; and based on the historical use behavior data, obtaining the loss risk level of the user, generating a member maintenance strategy for the user based on the payment capability level and the loss risk level of the user; and based on the member maintenance strategy, executing a fee deduction operation of subscribing a member service on a platform account of the user. By adopting the method, the accurate personalized member maintenance strategy can be customized for the user, and the user experience and the member retention rate are improved.

The invention relates to the technical field of artificial intelligence, and discloses a text data preprocessing method suitable for a large financial model, and the method comprises the steps: receiving an original financial text sequence sorted according to timestamps; performing word segmentation and part-of-speech tagging; dynamically calculating an emotional fluctuation index of each text unit based on the sliding time window; self-adaptively setting an emotion noise judgment threshold value and marking candidate noise units; semantic stability is evaluated through embedding space cosine similarity, and noise is confirmed for the second time; after the confirmation noise is removed, anaphora resolution and logic connection reconstruction are executed to repair context coherence; and outputting the pure text sequence. The system comprises corresponding function modules. According to the method, through a dynamic threshold and semantic dual verification mechanism, the effective information retention rate and the noise elimination accuracy rate are remarkably improved in a high-noise scene, and the quality and the processing throughput of financial large model input data are guaranteed.

The invention belongs to the technical field of internet information recommendation, and discloses a collaborative filtering recommendation method and device based on user grading, and the method comprises the following steps: S1, user grading processing: obtaining user data of a target platform, grading users based on a preset grading dimension, obtaining at least two user levels and a level weight corresponding to each user level; s2, data preprocessing and feature extraction: collecting user behavior data and content data, performing data cleaning, constructing a user-content interaction matrix, and extracting user features and content features; according to the scheme, the user level weights are deeply fused into similarity calculation and recommendation priority ranking, so that the preferences of high-value users and payment users can be remarkably amplified, and therefore, a more accurate high-quality recommendation result which better conforms to the identities and requirements of the users is obtained, and the satisfaction degree, retention rate and commercial conversion efficiency of core users are effectively improved.

The present application provides a short play user behavior analysis method and system, relating to the technical field of data processing, the method comprising: step 1, obtaining the dynamic behavior data of the user on the short play platform through multi-dimensional data, including real-time watching behavior data and advertisement interaction data, and establishing a user behavior time sequence database; step 2, based on the user behavior time sequence database, extracting three fixed detection points, constructing a user behavior feature topology relationship with the three fixed detection points as the vertex, and analyzing the user behavior feature topology relationship to generate a behavior feature adjustment value; step 3, based on the behavior feature adjustment value, dynamically identifying the user's preference in entertainment, information, socialization and emotional needs, while constructing a psychological resistance evaluation mechanism to quantify the user's psychological resistance to advertisements, and dynamically updating the resistance level label. The present application understands the user's preference and watching habit, helps the short play platform to adjust the content creation, and improves the user retention rate and the platform business value.

The invention discloses an interactive audience participation feedback method based on sentiment analysis and behavior prediction. User behavior data and text sentiment information are collected in real time through an audience terminal, behavior analysis and sentiment recognition are executed in parallel on a cloud server, and behavior tags and multi-level sentiment scores are generated; in combination with the historical portrait of the user, personalized interactive feedback content including an expression package, a knowledge point card, a microphone connection invitation or a voting entry is dynamically generated through a weighted scoring mechanism, and the pushing frequency is controlled to be not more than once every three minutes to avoid interference; meanwhile, when the individual continuously carries out negative feedback or the overall negative emotion ratio of the live broadcast room exceeds a threshold value, an instruction is automatically sent to the live broadcast stream pushing end, and real-time content adjustment is achieved. The method effectively solves the problems that traditional live broadcast interaction lags behind, is single in form and lacks emotional response, remarkably improves the audience immersion, participation and platform retention rate, and is suitable for various live broadcast scenes such as education, entertainment and e-commerce.

The invention discloses a user purchase willingness analysis method and system based on multi-source data. The method comprises the following steps: collecting and storing user behavior data, merchant behavior data and interactive language data in an e-commerce live broadcast process; constructing a BiGRU-ATT fusion model, training the BiGRU-ATT fusion model by using the collected interactive language data, and performing sentiment classification prediction on the newly generated interactive language data by the trained BiGRU-ATT fusion model to obtain a sentiment index EI; defining a user behavior sequence, calculating a behavior entropy weight and a retention rate of each stage in the user behavior sequence and an average time consumption ratio of the stage according to the collected user behavior data and merchant behavior data, and obtaining a behavior sequence index BI; and obtaining a fusion index CI indicating the purchase willingness of the user through the emotion index EI and the behavior sequence index BI. Therefore, the front emotion bullet screen can be guided, the overall atmosphere of the live broadcast room is improved, the user participation degree and satisfaction degree can be improved, and a new view angle is provided for live broadcast e-commerce marketing.

Embodiments of the present application disclose information content recommendation method and device, computer equipment and storage medium. The method comprises: setting the proportion value of the feature classification in the user portrait in the information recommendation, and setting the latest information proportion value; filtering information content meeting a first requirement by using the most important feature label in the user portrait, filtering information content meeting a second requirement by using the behavior data label in the user portrait; filtering information content meeting a third requirement by using the basic attribute data label in the user portrait; filtering information content meeting a fourth requirement according to the information publishing time; performing a merging operation to obtain an operation result; sorting the operation result; and sending the sorted result to a terminal. The method can improve the demand hit rate of users viewing information content, reduce the information cocoon house with smaller and smaller recommended information range, and improve the activity rate and retention rate of users.

The invention discloses an intelligent hierarchical management method and system based on customer full life cycle value evaluation, and relates to the field of big data processing, and the method comprises the steps: forming an initial data set through a multi-source heterogeneous data collection terminal and a preprocessing module, and extracting feature data in stages; a dynamic weight adjustment model is adopted to quantify the customer value of each stage, and a comprehensive score is calculated by combining value, stability and developability indexes; realizing customer clustering layering according to the grading threshold value by using an improved K-means algorithm; aiming at customer characteristics of different grades and stages, matching corresponding strategies in a personalized management strategy library; the strategy effect is evaluated through indexes such as the value improvement rate and the retention rate, analysis and optimization are carried out when the standard is not reached, and continuous monitoring is carried out after the standard is reached. The method has the advantages that through full-life-cycle multi-source data integration and dynamic value evaluation, accurate classification and personalized strategy matching of customers are achieved, closed-loop optimization is formed, the customer operation accuracy and value conversion are efficiently improved, and the management cost is reduced.

The invention relates to the technical field of big data and the technical field of artificial intelligence, and discloses a retention rate prediction method, device and equipment based on a large model, and the method comprises the steps: carrying out the feature extraction of an emotion description text corresponding to an emotion level, obtaining the emotion features of an IPTV account, carrying out the fusion of the behavior features of the IPTV account and the emotion features of the IPTV account, and carrying out the prediction of the retention rate of the IPTV account. The fusion features are utilized to retrieve similar cases in a vector knowledge base, and context information corresponding to the similar cases is obtained; the fusion features, the context information corresponding to the similar cases and the multiple time intervals are written into a prompt word template, a prompt text of the IPTV account is obtained, the prompt text of the IPTV account is processed through a large model, a prediction result of the IPTV account is obtained, and the retention rate of the IPTV account in each time interval is obtained from the prediction result of the IPTV account. The acquisition efficiency of the retention rate can be improved.

The invention provides a data processing method and device, and the method supports the inputting of a commodity screening demand of a natural language according to an actual live broadcast demand during the screening of a commodity in a live broadcast session, and obtains a commodity screening tag used for describing the commodity screening demand based on an intelligent analysis model. And searching and pushing a target commodity with the commodity screening label in a commodity library. The intelligent analysis model can understand the meaning of the natural language, so that the commodity screening demand of inputting the natural language is supported, the mode of inputting the commodity screening demand is free and not fixed, and an anchor or a merchant supporting a live broadcast session can express the commodity screening demand more conveniently. According to the method and the device, commodity screening based on multiple labels is supported, so that the input of commodity screening requirements with complex expression can be supported, the commodity screening efficiency and accuracy can be improved, the viscidity and retention rate of a user to a live broadcast platform, a live broadcast room or a live broadcast session in the live broadcast room can be improved, and the conversion rate of commodities can be improved.

The invention discloses an advertisement design and delivery management method based on artificial intelligence, and belongs to the technical field of artificial intelligence. According to the method, through full-process closed loop of multi-dimensional quality vector initialization, real-time cognitive impedance calculation, joint attenuation model construction, composite reward function design and double-layer agent decision, dynamic optimization of advertisement originality and fine management of user experience are realized, and the intelligent level of advertisement putting and the user retention rate are remarkably improved; by collecting implicit behaviors of a user in real time and combining pre-estimation of cognitive load and novelty entropy, user granularity quantification of cognitive impedance intensity is achieved, an eye movement model and behavior data are fused, and whether psychological resistance is generated by the user due to too high creative complexity or not is accurately judged through a sight escape speed mean value and a page residence time fluctuation coefficient; a dynamic threshold mechanism ensures that impedance judgment adapts to different user groups, and misjudgment is avoided.

The application relates to the technical field of artificial intelligence, and discloses a text data preprocessing method suitable for a financial large model, which comprises the following steps: receiving an original financial text sequence sorted by time stamp; performing word segmentation and part-of-speech tagging; dynamically calculating an emotional fluctuation index of each text unit based on a sliding time window; adaptively setting an emotional noise judgment threshold and marking a candidate noise unit; evaluating semantic stability through embedding space cosine similarity, and secondarily confirming noise; after the confirmed noise is removed, performing reference resolution and logical connection reconstruction to repair context coherence; and outputting a pure text sequence. The system comprises corresponding function modules. Through the dynamic threshold and semantic double verification mechanism, the application significantly improves the effective information retention rate and noise removal accuracy in a high-noise scene, and guarantees the quality and processing throughput of input data of the financial large model.

The invention discloses an information recommendation method and device, electronic equipment and a readable storage medium, and belongs to the technical field of information recommendation. User data of all users in a game is obtained; predicting game interaction compactness between the two users based on the user data of any two users; predicting a target game interaction behavior matched with the two users based on the game interaction closeness; in the embodiment of the invention, the interaction guide information is generated based on the target game interaction behavior and recommended to the two users to guide the two users to execute the target game interaction behavior, so that the stickiness between players can be improved, and the retention rate of the players is improved.

The invention provides an advertisement putting method and device and electronic equipment, and the method comprises the steps: obtaining attribute parameters of a target user in response to an advertisement putting request corresponding to the target user; the attribute parameter is at least used for indicating the acceptance degree of the target user to the advertisement with the quality parameter meeting the preset condition; the attribute parameter is determined based on a specified feedback behavior of the target user for the historical advertisement; the quality parameter is determined by an image quality evaluation model and a specified feedback behavior for the to-be-delivered advertisement; acquiring a plurality of to-be-delivered advertisements from a preset advertisement database; the to-be-delivered advertisement has a corresponding quality parameter; and based on the attribute parameter of the target user and the quality parameters of the plurality of to-be-delivered advertisements, determining a target advertisement, and delivering the target advertisement to the target user. According to the method, when the advertisement put to the user is determined, the receiving degree of the low-quality advertisement by the user is considered, the user experience is improved, and the retention rate of the user is improved to a certain extent.

Systems and methods for providing item recommendations to increase retention rates of customers are disclosed. In some embodiments, a disclosed method includes: receiving, from a computing device, a recommendation request for recommending items to a customer; determining, based on the recommendation request, an anchor item to be displayed to the customer via a user interface; determining relevance scores between the anchor item and a plurality of candidate items; determining retention scores between the anchor item and the plurality of candidate items; generating, using at least one machine learning model, a ranked list of recommended items based on the plurality of candidate items, the relevance scores and the retention scores; and transmitting to the computing device the ranked list of recommended items to be displayed to the customer with the anchor item on the user interface.

The invention discloses a virtual item obtaining method, device and equipment and a computer readable storage medium, and belongs to the technical field of computers. The method comprises the steps that a first virtual article collected by a first object in a virtual battlefield and a second virtual article collected by a second object in the virtual battlefield are displayed, and the first object and the second object belong to the same game team; based on triggering of a virtual item distribution function, displaying a third virtual item to be acquired by the first object, the third virtual item being determined according to the first virtual item and the second virtual item; and obtaining the third virtual article. According to the method, the obtaining mode of the virtual item is more diversified, the flexibility is higher, and the user retention rate and the user participation degree of the game are further improved.