A movie comprehensive scoring method and system based on dynamic weights

Abstract

The application relates to the technical field of film scoring, and discloses a film comprehensive scoring method and system based on dynamic weights, which aims to solve the problem of distorted scoring results in the prior art, and mainly comprises the following steps: obtaining heat-related data and quality-related data of a film; extracting heat feature indexes based on a time sequence of a play quantity and generating a comprehensive heat score; judging the data integrity of multiple sub-dimensions of a quality score, and when data is missing, adaptively redistributing the weights of the remaining sub-dimensions to generate a quality score; extracting a film type feature, and dynamically calculating the fusion weights of the quality score and the heat score in the comprehensive score based on the type feature, heat change features, and the relative levels of the quality score and the heat score; and applying the dynamic weights to the quality score and the comprehensive heat score for weighted fusion to generate a final comprehensive score which is automatically adjusted according to the state of the film. The application improves the accuracy, robustness and timeliness of the scoring.

Description

Technical Field

[0001] This invention relates to the field of film rating technology, specifically to a comprehensive film rating method and system based on dynamic weights. Background Technology

[0002] In the field of film content evaluation and recommendation, current comprehensive scoring systems typically employ a multi-dimensional index weighted fusion technical solution. Specifically, the system independently scores the film's content quality, market popularity, and other dimensions, and then linearly weights these scores using pre-set fixed weights or semi-fixed weights based on simple rules to calculate a comprehensive score result, which is then used for subsequent ranking, recommendation, or retrieval services.

[0003] However, the aforementioned existing technical solutions have significant drawbacks in practical applications: the comprehensive scoring results lack adaptability to changes in the film's status and are difficult to differentiate for films at different quality levels, different lifecycle stages, or different data integrity conditions. For example, a high-quality film that is still in its initial launch phase may have an extremely low popularity score, and its comprehensive score will be severely underestimated under static weighting, resulting in its content value not being reasonably reflected; conversely, a film that has become extremely popular but is of mediocre quality may have its score overestimated due to its excessive popularity, causing the scoring results to be mismatched with the film's actual content value or long-term market performance.

[0004] The technical reasons for the aforementioned deficiencies lie in the fact that existing solutions generally employ static weights or fixed-rule weight fusion methods, with their scoring logic remaining at a simple numerical summation level. Specifically, existing technologies neither model and analyze the state correlation between quality scores and popularity scores, nor incorporate key factors such as popularity trends, the reliability of quality data, and differences in film genres into the weight adjustment logic. When film popularity changes rapidly or some quality data (such as third-party ratings) is missing, the lack of a mechanism to judge data integrity and the ability to perceive the dynamic state of the film easily leads to distorted comprehensive scoring results, and insufficient stability and reliability of the scores, ultimately affecting the overall performance of content ranking or recommendation systems. Summary of the Invention

[0005] This invention aims to solve the problem that existing film rating methods cannot perceive the film's status, resulting in distorted comprehensive rating results. It proposes a film comprehensive rating method and system based on dynamic weights.

[0006] The technical solution adopted by the present invention to solve the above-mentioned technical problems is as follows: In a first aspect, the present invention provides a film comprehensive scoring method based on dynamic weights, the method comprising: Obtain basic data for the film to be evaluated, including popularity-related data and quality-related data; Based on the changes in the popularity-related data at multiple time points within a preset time window, a structured feature analysis is performed to extract multiple popularity feature indicators that reflect the popularity status of the film. Based on each popularity feature indicator, a weighted combination is performed according to a preset combination strategy to generate a comprehensive popularity score for the film to be evaluated. The quality score of the film to be evaluated is calculated based on the quality-related data. Data integrity and availability are judged for each of the multiple quality sub-dimensions that constitute the quality score. When data missing or unavailable for any quality sub-dimension is detected, the weight ratio of the remaining available quality sub-dimensions is adaptively reallocated, and the quality score of the film to be evaluated is generated based on the reallocated weight ratio. Extract the film type features of the film to be evaluated, and perform joint analysis based on the film type features, the popularity feature indicators reflecting the popularity change features, and the relative level between the quality score and the overall popularity score. Dynamically calculate the fusion weight of the quality score and the overall popularity score in the final comprehensive evaluation. The dynamically calculated fusion weights are applied to the quality score and the overall popularity score respectively for weighted fusion to generate the final comprehensive score result of the film to be evaluated. The final comprehensive score result is automatically adjusted as the film status changes.

[0007] This invention acquires basic film data, extracts popularity features, adaptively calculates quality scores, dynamically determines fusion weights, and generates a comprehensive score, enabling the scoring results to automatically adjust according to the film's status. Compared to existing static weighting techniques, this invention significantly improves the scoring's adaptability to different film types, lifecycle stages, and data integrity conditions. It solves the problems of underestimating high-quality films, overestimating popular films, and data gaps causing score distortion, thus improving the accuracy and reliability of the scoring.

[0008] Furthermore, the heat characteristic indicators include two or more of the following: Stability characteristics used to reflect the fluctuations in the number of views of the film to be evaluated; Growth characteristics used to reflect trends in play count; A characteristic used to reflect the duration of heat intensity; Peak characteristics used to reflect peak performance in terms of play count.

[0009] This invention performs a structured analysis of film popularity by extracting multi-dimensional popularity features such as stability, growth, persistence, and peak performance of play counts, overcoming the limitations of a single popularity indicator. These features comprehensively depict the film's market performance, providing more accurate state input for dynamic weight adjustments and making the popularity score more closely reflect the film's actual popularity changes.

[0010] Furthermore, the multiple quality sub-dimensions that constitute a film's quality rating include combinations of the following: A third-party rating dimension is formed based on third-party evaluation information of the film to be evaluated; The picture quality dimension is based on the maximum available resolution or playback quality of the video. Content completeness dimension is formed based on the completeness of the film's metadata.

[0011] This invention decomposes quality scoring into multiple sub-dimensions, including third-party ratings, picture quality, and content completeness, constructing a multi-dimensional quality assessment system. This avoids the bias of relying on a single quality indicator, making the evaluation of film content value more comprehensive and objective, and laying a solid foundation for the reasonable integration of subsequent quality scores and overall popularity scores.

[0012] Furthermore, the adaptive reallocation of the weight proportions of the remaining available quality sub-dimensions includes: Based on the film type of the film to be evaluated, retrieve the basic preset weight ratio of each quality sub-dimension corresponding to the film type; When a missing data is detected in a certain target quality sub-dimension, the basic preset weight ratio of that target quality sub-dimension is stripped, and the stripped weight ratio is compensated and distributed to the remaining available quality sub-dimensions according to preset rules, so that the sum of the weight ratios of the redistributed available quality sub-dimensions meets the calculation requirements.

[0013] This invention introduces an adaptive weight allocation mechanism into the quality scoring process, automatically reallocating weights to available dimensions when data for a certain sub-dimension is missing. This ensures stable calculation of quality scores even in scenarios with incomplete data, significantly improving the robustness and practicality of the scoring system and preventing overall evaluation failure due to missing local data.

[0014] Furthermore, the dynamic calculation of the fusion weight of the quality score and the overall popularity score in the final comprehensive evaluation includes: Based on the analysis of the relative levels between the quality score and the overall popularity score, when the film to be evaluated is in a state where the quality score is higher than the preset quality threshold and the overall popularity score is lower than the preset popularity threshold, the weight of the quality score in the final overall score result is increased, and the weight of the overall popularity score is reduced accordingly.

[0015] This invention dynamically adjusts the weighting of quality scores and overall popularity scores based on their relative levels: when a film is in a high-quality, low-popularity state, the weighting of the quality score is automatically increased. This effectively protects the reasonableness of ratings for high-quality, less popular films, allowing their content value to stand out and improving the healthy development of the content ecosystem. Furthermore, the dynamic calculation of the fusion weight of the quality score and the overall popularity score in the final comprehensive evaluation also includes: Based on the analysis of the popularity change characteristics reflected by the popularity feature indicators, when the film to be evaluated is in a state of rapid popularity growth with a playback rate higher than a preset speed threshold, the weight of the comprehensive popularity score in the final comprehensive score result is increased.

[0016] This invention dynamically adjusts weights based on trends in popularity: when a film's popularity grows rapidly, the weight of the overall popularity score is automatically increased. This allows the score to respond promptly to changes in market popularity, avoiding the problem of the score lagging behind the film's actual popularity, and improving the timeliness and sensitivity of the score.

[0017] Furthermore, after extracting multiple popularity feature indicators based on the popularity-related data, and before generating a comprehensive popularity score, the method further includes: The extracted heat feature indicators of different dimensions or units are uniformly normalized.

[0018] This invention eliminates the impact of dimensional differences on weighted combinations by normalizing popularity characteristic indicators with different dimensions. This ensures the mathematical consistency of popularity score calculation, provides standardized input for subsequent weight fusion, and improves the scientific rigor and accuracy of the scoring process.

[0019] Furthermore, the popularity-related data in the basic data includes: unique content identifier, unified time span, most recent update time, and play count change data collected within a continuous observation time window; The quality-related data includes: the number of tags, the number of cast and crew information, the highest supported resolution parameters, and third-party rating values.

[0020] This invention clarifies the specific content of basic data collection, including playback volume time series, metadata, and resolution, providing comprehensive and standardized data support for various scoring calculations. It ensures the feasibility of the method and data consistency, facilitating its widespread application in large-scale film libraries.

[0021] Furthermore, the method also includes: Based on the final comprehensive score, the films are sorted, recommended, or retrieved.

[0022] This invention uses the final comprehensive score for film ranking, recommendation, or retrieval, directly translating improvements in the scoring method into performance enhancements for the content distribution system. Ranking and recommendation based on more accurate scores improves user experience, enhances recommendation effectiveness, and effectively connects technological and business value.

[0023] Secondly, the present invention provides a film comprehensive scoring system based on dynamic weights, for implementing the film comprehensive scoring method based on dynamic weights as described in the first aspect, the system comprising: The data acquisition module is used to acquire basic data of the film to be evaluated, including popularity-related data and quality-related data. The comprehensive popularity score generation module is used to perform structured feature analysis on the changes of the popularity-related data at multiple time points within a preset time window, extract multiple popularity feature indicators that reflect the popularity status of the film, and generate a comprehensive popularity score of the film to be evaluated based on the weighted combination of each popularity feature indicator according to a preset combination strategy. The quality score generation module is used to calculate the quality score of the film to be evaluated based on the quality-related data, and to perform data integrity and availability judgments on the multiple quality sub-dimensions that constitute the quality score. When the data corresponding to any quality sub-dimension is detected to be missing or unavailable, the weight ratio of the remaining available quality sub-dimensions is adaptively reallocated, and the quality score of the film to be evaluated is calculated and generated based on the reallocated weight ratio. The fusion weight calculation module is used to extract the film type features of the film to be evaluated, and to perform joint analysis based on the film type features, the popularity feature indicators reflecting the popularity change features, and the relative level between the quality score and the comprehensive popularity score, and dynamically calculate the fusion weight of the quality score and the comprehensive popularity score in the final comprehensive evaluation. The final comprehensive score generation module is used to apply the dynamically calculated fusion weights to the quality score and the comprehensive popularity score respectively for weighted fusion, and generate the final comprehensive score result of the film to be evaluated. The final comprehensive score result is automatically adjusted as the film status changes.

[0024] This invention automates the execution of the aforementioned methods through the collaborative work of modules such as data acquisition, popularity score generation, quality score generation, fusion weight calculation, and comprehensive score generation. The system can be directly deployed on a server or in the cloud, providing standardized film value assessment capabilities for video platforms and recommendation systems, and has promising prospects for industrial applications.

[0025] The beneficial effects of this invention are as follows: The dynamic weight-based comprehensive film scoring method and system provided by this invention, by introducing a film status awareness mechanism, solves the shortcomings of existing static weighted scoring techniques that cannot adapt to changes in film quality levels, popularity stages, and data integrity. This invention first performs structured analysis on playback data to accurately depict popularity status. Simultaneously, it introduces adaptive weight allocation in quality scoring to address scenarios with missing data. Then, it dynamically determines the fusion weights based on the relative relationship between film type, quality, and popularity, enabling the comprehensive score to automatically adjust according to the film's status. This significantly improves the accuracy, robustness, and timeliness of the scoring, ensuring that high-quality films are not underestimated and that popular films reflect their value promptly. Ultimately, it provides a more reasonable evaluation basis for film ranking, recommendation, and retrieval systems, effectively improving user experience and platform content management efficiency. Attached Figure Description

[0026] Figure 1 A flowchart illustrating a dynamic weight-based comprehensive film scoring method provided for an embodiment; Figure 2 A flowchart illustrating another dynamic weight-based comprehensive film scoring method provided for an embodiment; Figure 3 This is a schematic diagram of the structure of a film comprehensive scoring system based on dynamic weights provided for an embodiment. Detailed Implementation

[0027] Existing film scoring technologies employ static weighting to fuse multi-dimensional scores, which fails to perceive and adapt to changes in a film's status under varying quality levels, popularity stages, and data integrity conditions, leading to distorted scoring results. Therefore, overcoming the technical shortcomings of existing technologies—namely, poor adaptability of scoring results due to static weighting and the inability to dynamically adjust to the actual state of the film—has become a pressing technical problem to be solved in this field.

[0028] Based on this, the technical solution of this invention is proposed. In this invention, firstly, basic data of the film is acquired; then, through structured analysis of the playback volume time series, popularity feature indicators are extracted and a comprehensive popularity score is generated. Simultaneously, quality-related data is assessed for data integrity by sub-dimension and weights are adaptively assigned to generate a quality score. Next, film type features are extracted and combined with popularity change features and the relative levels of the quality score and comprehensive popularity score for joint analysis, dynamically calculating their fusion weight in the final comprehensive score. Finally, the dynamic weights are applied to the quality score and comprehensive popularity score for weighted fusion, generating a final comprehensive score result that automatically adjusts with changes in the film's status.

[0029] The technical solutions in this embodiment will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0030] Please see Figure 1 and Figure 2 The film comprehensive scoring method based on dynamic weights provided in this embodiment includes the following steps: Step 1: Basic Data Collection Obtain basic data for the film to be evaluated, including popularity-related data and quality-related data.

[0031] Specifically, this step aims to provide comprehensive data input for subsequent film status analysis. In practical applications, the first step is to acquire basic data about the film to be evaluated. The collected data includes at least the following two categories: first, film popularity-related data, mainly referring to the changes in play counts collected within a preset continuous time window (e.g., multiple consecutive days), used to reflect the film's market performance; second, film quality-related data, specifically including the completeness of film metadata (e.g., number of tags, number of cast and crew information), available maximum resolution information, and the existence of available third-party rating data. Both types of data include basic film information, such as unique content identifiers, film titles, statistical time spans, and the most recent update time.

[0032] The data collection provided above lays the foundation for subsequent features extraction, quality score calculation, and dynamic weight determination.

[0033] Step 2, Heat Feature Extraction: Based on the changes in the popularity-related data at multiple time points within a preset time window, structured feature analysis is performed to extract multiple popularity feature indicators that reflect the popularity status of the film. Based on each popularity feature indicator, a weighted combination is performed according to a preset combination strategy to generate a comprehensive popularity score for the film to be evaluated.

[0034] Specifically, this step aims to perform structured analysis on the film's playback data to quantify its market popularity. In practical applications, based on the playback data collected in step 1 within a preset continuous time window, statistical analysis is performed on its fluctuations at multiple time points to extract multi-dimensional feature indicators that reflect the film's popularity.

[0035] In this embodiment, the popularity characteristic indicators include the following four items: first, stability characteristics reflecting the fluctuation of play volume, measuring the stability of popularity; second, growth characteristics reflecting the trend of play volume changes, measuring the rising or falling trend of popularity; third, persistence characteristics reflecting the duration of popularity performance, measuring the ability to maintain a high-heat state; and fourth, peak characteristics reflecting the peak performance of play volume, measuring the explosive intensity of popularity.

[0036] The aforementioned popularity indicators can be obtained through statistical analysis or trend analysis. After extraction, each popularity indicator is normalized to eliminate dimensional differences, and then weighted according to a preset fusion strategy to generate a comprehensive popularity score for the film to be evaluated. The weighting ratio in the preset fusion strategy can be set according to actual business needs, providing standardized popularity status input for subsequent dynamic weight adjustments.

[0037] Step 3: Quality Score Calculation and Adaptive Weight Allocation: The quality score of the film to be evaluated is calculated based on the quality-related data. Data integrity and availability are judged for each of the multiple quality sub-dimensions that constitute the quality score. When data missing or unavailable for any quality sub-dimension is detected, the weight ratio of the remaining available quality sub-dimensions is adaptively reallocated, and the quality score of the film to be evaluated is generated based on the reallocated weight ratio.

[0038] Specifically, this step aims to dynamically adjust the internal composition of the quality score based on data availability, ensuring the reliability of the quality assessment in scenarios with incomplete data. In practical applications, firstly, based on the quality-related data collected in step 1, a quality score is calculated for the film to be evaluated, and the quality score is decomposed into a combination of multiple quality sub-dimensions. In this embodiment, the quality sub-dimensions include the following three items: first, a third-party rating dimension based on third-party evaluation information, reflecting external professional evaluation; second, a picture quality dimension based on the film's maximum available resolution or playback quality, reflecting the level of technical specifications; and third, a content completeness dimension based on the completeness of the film's metadata (such as the number of tags, the number of cast and crew information, etc.), reflecting the richness of the content.

[0039] During the calculation process, data integrity and availability are assessed for each of the aforementioned quality sub-dimensions. When data loss or unavailability is detected for any quality sub-dimension, an adaptive weight allocation mechanism is automatically activated. Specifically, the following steps are taken: First, based on the film type to be evaluated, the basic preset weight ratios for each quality sub-dimension corresponding to that type are retrieved. When data loss is detected for a target quality sub-dimension, the basic preset weight ratio for that target quality sub-dimension is stripped, and the stripped weight ratio is redistributed to the remaining available quality sub-dimensions according to preset rules (e.g., allocated according to the original ratio or evenly). After redistribution, the sum of the weight ratios of all available quality sub-dimensions satisfies the normalization calculation requirements. The available quality sub-dimensions are then weighted according to the redistributed weight ratios to generate the final quality score for the film to be evaluated. Through this adaptive mechanism, this embodiment avoids the problem of overall quality assessment distortion caused by the loss of a single quality data point, ensuring the stability and reliability of the quality score under different data conditions, and providing a reliable input for content value assessment for subsequent dynamic weight fusion.

[0040] Step 4: Determine dynamic weights: The film type features of the film to be evaluated are extracted, and a joint analysis is performed based on the film type features, the popularity feature indicators reflecting the popularity change features, and the relative level between the quality score and the overall popularity score. The fusion weight of the quality score and the overall popularity score in the final comprehensive evaluation is dynamically calculated.

[0041] Specifically, this step aims to adaptively calculate the fusion weight of quality score and comprehensive popularity score in the final comprehensive evaluation based on the joint analysis of the film's multi-dimensional status, so that the weight allocation can be dynamically adjusted as the film's status changes.

[0042] In practical applications, the film type characteristics of the film to be evaluated are first extracted as one of the bases for weight adjustment. Then, based on the popularity feature indicators generated in step 2, the film's popularity change characteristics (e.g., the trend of view growth) are obtained, and the relative level between the comprehensive popularity score generated in step 2 and the quality score generated in step 3 is obtained (e.g., the quality score is higher than the comprehensive popularity score, or the comprehensive popularity score is higher than the quality score). The above film type characteristics, popularity change characteristics, and the relative level of the quality score and the comprehensive popularity score are jointly analyzed to dynamically calculate the fusion weight of the two in the final comprehensive evaluation.

[0043] In this embodiment, dynamic weight calculation includes the following two typical adjustment mechanisms: First, adjustments are made based on the relative level of the situation. When the analysis finds that the film to be evaluated is in a high-quality, low-popularity state where the quality score is higher than the preset quality threshold and the overall popularity score is lower than the preset popularity threshold, the weight of the quality score in the final overall score result is increased, and the weight of the overall popularity score is reduced accordingly, so as to protect high-quality content from being underestimated.

[0044] Secondly, adjustments are made based on the characteristics of popularity changes. When the analysis finds that the film to be evaluated is in a state of rapid growth in popularity where the rate of increase in play count is higher than the preset speed threshold, the weight of the comprehensive popularity score in the final comprehensive score result is increased to reflect the rapid increase in the film's popularity.

[0045] The aforementioned weight adjustment method can be implemented based on preset rule strategies, mapping relationship tables, or other feasible methods. Through the above mechanism, this embodiment achieves adaptive changes in fusion weights according to the film's state, providing an optimal weight configuration that conforms to the film's current actual state for subsequent comprehensive score generation.

[0046] Step 5: Generation and Output of Comprehensive Score The dynamically calculated fusion weights are applied to the quality score and the overall popularity score respectively for weighted fusion to generate the final comprehensive score result of the film to be evaluated. The final comprehensive score result is automatically adjusted as the film status changes.

[0047] Specifically, this step aims to apply the dynamic weights calculated in the previous steps to the film's quality score and popularity score, generating a final comprehensive score that reflects the film's real-time status.

[0048] In practical applications, the fusion weights dynamically calculated in step 4 are assigned to the quality score generated in step 3 and the comprehensive popularity score generated in step 2, respectively, and then weighted and fused. Since the fusion weights are derived from a joint analysis of film type characteristics, popularity change characteristics, and the relative levels of the quality score and the comprehensive popularity score, the final comprehensive score can automatically adjust according to changes in the film's status: when the film is in a high-quality, low-popularity state, the quality score carries a higher weight, and the comprehensive score better reflects its content value; when the film is experiencing rapid popularity growth, the comprehensive popularity score carries a higher weight, and the comprehensive score can promptly reflect its market performance. This achieves adaptive generation of the film's comprehensive score.

[0049] In this embodiment, the final comprehensive score result can be used for subsequent film ranking, recommendation, or retrieval processing, providing the platform with an evaluation basis that better reflects the actual value of the films. Specifically, after generating the final comprehensive score result for the films to be evaluated in step 5, this score is used as a unified basis for film value assessment and applied to multiple business scenarios: In the ranking scenario, multiple films can be sorted in descending order based on the final comprehensive score to generate a film list for ranking display or content filtering; in the recommendation scenario, the final comprehensive score is used as an important feature or ranking basis for the recommendation algorithm, prioritizing the push of films with higher scores to users; in the retrieval scenario, relevant films can be retrieved based on query conditions, and the retrieval results can be re-ranked based on the final comprehensive score to improve the relevance of the results and user experience. Through the above applications, this embodiment transforms the adaptive comprehensive score generated in the preceding steps into actual business value output, enabling high-quality films to receive reasonable exposure and hot films to reflect their popularity in a timely manner, thereby improving the platform's content distribution efficiency and user satisfaction.

[0050] Based on the above technical solution, this embodiment also provides a dynamic weight-based comprehensive film scoring system to implement the dynamic weight-based comprehensive film scoring method described in the embodiment. Please refer to [link to relevant documentation]. Figure 2 The system includes: The data acquisition module is used to acquire basic data of the film to be evaluated, including popularity-related data and quality-related data. The comprehensive popularity score generation module is used to perform structured feature analysis on the changes of the popularity-related data at multiple time points within a preset time window, extract multiple popularity feature indicators that reflect the popularity status of the film, and generate a comprehensive popularity score of the film to be evaluated based on the weighted combination of each popularity feature indicator according to a preset combination strategy. The quality score generation module is used to calculate the quality score of the film to be evaluated based on the quality-related data, and to perform data integrity and availability judgments on the multiple quality sub-dimensions that constitute the quality score. When the data corresponding to any quality sub-dimension is detected to be missing or unavailable, the weight ratio of the remaining available quality sub-dimensions is adaptively reallocated, and the quality score of the film to be evaluated is calculated and generated based on the reallocated weight ratio. The fusion weight calculation module is used to extract the film type features of the film to be evaluated, and to perform joint analysis based on the film type features, the popularity feature indicators reflecting the popularity change features, and the relative level between the quality score and the comprehensive popularity score, and dynamically calculate the fusion weight of the quality score and the comprehensive popularity score in the final comprehensive evaluation. The final comprehensive score generation module is used to apply the dynamically calculated fusion weights to the quality score and the comprehensive popularity score respectively for weighted fusion, and generate the final comprehensive score result of the film to be evaluated. The final comprehensive score result is automatically adjusted as the film status changes.

[0051] It is understood that since the dynamic weight-based film comprehensive scoring system described in this embodiment is a system for implementing the dynamic weight-based film comprehensive scoring method described in the embodiment, the system disclosed in the embodiment is relatively simple to describe because it corresponds to the method disclosed in the embodiment. For relevant parts, please refer to the description of the method, and it will not be repeated here.

Claims

1. A film comprehensive scoring method based on dynamic weights, characterized in that, The method includes: Obtain basic data for the film to be evaluated, including popularity-related data and quality-related data; Based on the changes in the popularity-related data at multiple time points within a preset time window, a structured feature analysis is performed to extract multiple popularity feature indicators that reflect the popularity status of the film. Based on each popularity feature indicator, a weighted combination is performed according to a preset combination strategy to generate a comprehensive popularity score for the film to be evaluated. The quality score of the film to be evaluated is calculated based on the quality-related data. Data integrity and availability are judged for each of the multiple quality sub-dimensions that constitute the quality score. When data missing or unavailable for any quality sub-dimension is detected, the weight ratio of the remaining available quality sub-dimensions is adaptively reallocated, and the quality score of the film to be evaluated is generated based on the reallocated weight ratio. Extract the film type features of the film to be evaluated, and perform joint analysis based on the film type features, the popularity feature indicators reflecting the popularity change features, and the relative level between the quality score and the overall popularity score. Dynamically calculate the fusion weight of the quality score and the overall popularity score in the final comprehensive evaluation. The dynamically calculated fusion weights are applied to the quality score and the overall popularity score respectively for weighted fusion to generate the final comprehensive score result of the film to be evaluated. The final comprehensive score result is automatically adjusted as the film status changes.

2. The film comprehensive scoring method based on dynamic weights according to claim 1, characterized in that, The heat characteristic indicators include two or more of the following: Stability characteristics used to reflect the fluctuations in the number of views of the film to be evaluated; Growth characteristics used to reflect trends in play count; A characteristic used to reflect the duration of heat intensity; Peak characteristics used to reflect peak performance in terms of play count.

3. The film comprehensive scoring method based on dynamic weights according to claim 1, characterized in that, The multiple quality sub-dimensions that constitute a film's quality rating include a combination of the following: A third-party rating dimension is formed based on third-party evaluation information of the film to be evaluated; The picture quality dimension is based on the maximum available resolution or playback quality of the video. Content completeness dimension is formed based on the completeness of the film's metadata.

4. The film comprehensive scoring method based on dynamic weights according to claim 1 or 3, characterized in that, The adaptive reallocation of the weight proportions of the remaining available quality sub-dimensions includes: Based on the film type of the film to be evaluated, retrieve the basic preset weight ratio of each quality sub-dimension corresponding to the film type; When a missing data is detected in a certain target quality sub-dimension, the basic preset weight ratio of that target quality sub-dimension is stripped, and the stripped weight ratio is compensated and distributed to the remaining available quality sub-dimensions according to preset rules, so that the sum of the weight ratios of the redistributed available quality sub-dimensions meets the calculation requirements.

5. The film comprehensive scoring method based on dynamic weights according to claim 1, characterized in that, The dynamic calculation of the fusion weight of the quality score and the overall popularity score in the final comprehensive evaluation includes: Based on the analysis of the relative levels between the quality score and the overall popularity score, when the film to be evaluated is in a state where the quality score is higher than the preset quality threshold and the overall popularity score is lower than the preset popularity threshold, the weight of the quality score in the final overall score result is increased, and the weight of the overall popularity score is reduced accordingly.

6. The film comprehensive scoring method based on dynamic weights according to claim 1, characterized in that, The dynamic calculation of the fusion weight of the quality score and the overall popularity score in the final comprehensive evaluation also includes: Based on the analysis of the popularity change characteristics reflected by the popularity feature indicators, when the film to be evaluated is in a state of rapid popularity growth with a playback rate higher than a preset speed threshold, the weight of the comprehensive popularity score in the final comprehensive score result is increased.

7. The film comprehensive scoring method based on dynamic weights according to claim 1, characterized in that, After extracting multiple popularity feature indicators based on the popularity-related data, and before generating a comprehensive popularity score, the method further includes: The extracted heat feature indicators of different dimensions or units are uniformly normalized.

8. The film comprehensive scoring method based on dynamic weights according to claim 1, characterized in that, The popularity-related data in the basic data includes: unique content identifier, unified time span, most recent update time, and play count change data collected within a continuous observation time window; The quality-related data includes: the number of tags, the number of cast and crew information, the highest supported resolution parameters, and third-party rating values.

9. The film comprehensive scoring method based on dynamic weights according to claim 1, characterized in that, The method further includes: Based on the final comprehensive score, the films are sorted, recommended, or retrieved.

10. A film comprehensive scoring system based on dynamic weights, characterized in that, The system is used to implement the dynamic weight-based comprehensive film scoring method as described in any one of claims 1 to 9, the system comprising: The data acquisition module is used to acquire basic data of the film to be evaluated, including popularity-related data and quality-related data. The comprehensive popularity score generation module is used to perform structured feature analysis on the changes of the popularity-related data at multiple time points within a preset time window, extract multiple popularity feature indicators that reflect the popularity status of the film, and generate a comprehensive popularity score of the film to be evaluated based on the weighted combination of each popularity feature indicator according to a preset combination strategy. The quality score generation module is used to calculate the quality score of the film to be evaluated based on the quality-related data, and to perform data integrity and availability judgments on the multiple quality sub-dimensions that constitute the quality score. When the data corresponding to any quality sub-dimension is detected to be missing or unavailable, the weight ratio of the remaining available quality sub-dimensions is adaptively reallocated, and the quality score of the film to be evaluated is calculated and generated based on the reallocated weight ratio. The fusion weight calculation module is used to extract the film type features of the film to be evaluated, and to perform joint analysis based on the film type features, the popularity feature indicators reflecting the popularity change features, and the relative level between the quality score and the comprehensive popularity score, and dynamically calculate the fusion weight of the quality score and the comprehensive popularity score in the final comprehensive evaluation. The final comprehensive score generation module is used to apply the dynamically calculated fusion weights to the quality score and the comprehensive popularity score respectively for weighted fusion, and generate the final comprehensive score result of the film to be evaluated. The final comprehensive score result is automatically adjusted as the film status changes.

Images