SandBaseSandBase

检测视频虚假流量分析/Detect fake views in video

TikTokapi

About

检测视频虚假流量分析/Detect fake views in video

Documentation

检测视频虚假流量分析/Detect fake views in video

[中文]

用途:

  • 通过高级算法分析TikTok视频流量数据，精确检测可能存在的虚假观看量和不自然互动
  • 基于TikTok赛马机制(Traffic Pool)流量池理论，评估内容真实性和流量质量
  • 提供全面的欺诈风险分析，包含8种维度、20+指标的深度评估
  • 为创作者、MCN机构和内容管理者提供专业的流量质量报告和优化建议

参数:

  • item_id: 视频作品ID，必填参数，可从视频URL中提取(例如: https://www.tiktok.com/@tiktok/video/7460937381265411370 中的7460937381265411370)
  • content_category: 内容分类，可选参数，影响互动率基准值，选项包括:
    • default: 默认类别，通用内容
    • entertainment: 娱乐内容，预期有较高互动率
    • education: 教育内容，预期有适中互动和较高收藏率
    • product: 产品内容，预期有较低互动但较高转化
    • verified_large: 大型认证账号，预期互动率适当降低

返回内容详解:

  • video_metrics: 视频核心指标

    • total_views: 总观看量，视频被观看的总次数
    • total_likes: 总点赞数，用户点赞互动次数
    • total_comments: 总评论数，用户评论互动次数
    • total_favorites: 总收藏数，用户收藏次数
    • total_shares: 总分享数，用户分享次数
    • engagement_rates: 互动率指标，值越高越好
      • like_ratio: 点赞率，正常值 1-10%，大账号可能较低
      • comment_ratio: 评论率，正常值 0.1-0.5%，高于1%极佳
      • favorite_ratio: 收藏率，正常值 0.05-0.8%
      • share_ratio: 分享率，正常值 0.05-0.5%，高于1%极佳

  • creator_metrics: 创作者账号健康指标

    • account_age_days: 账号存在天数，越长越可信
    • follower_count: 粉丝数量，影响预期观看量
    • verified: 是否验证账号，认证账号可信度更高
    • trust_score: 账号信任度评分(0-100)，越高越可信

  • content_metrics: 内容质量指标

    • content_type: 内容类型(video, image等)
    • created_by_ai: 是否AI生成，AI生成内容可能有特定流量模式
    • high_quality_upload: 是否高质量上传，高质量上传更可信

  • fake_view_analysis: 虚假流量综合分析

    • fake_score: 虚假流量评分(0-100)，评分越低越好:
      • 0-20: 极低风险，自然流量模式
      • 20-40: 低风险，可能有少量异常但不构成问题
      • 40-60: 中等风险，存在值得关注的异常
      • 60-80: 高风险，明显的虚假流量特征
      • 80-100: 极高风险，几乎确定存在虚假流量
    • confidence_level: 风险等级，分为"Minimal", "Low", "Medium", "High"
    • estimated_fake_views: 估计虚假观看量，基于虚假流量模型推算
    • fake_view_percentage: 虚假观看百分比，虚假占总量的比例
    • is_suspicious: 是否可疑，综合判断是否需要关注
    • main_detection_reason: 主要检测原因，最显著的异常特征
    • component_scores: 各维度异常评分，各项都是0-100，越低越好:
      • engagement_score: 互动异常评分
      • distribution_score: 分布异常评分
      • consistency_score: 一致性异常评分
      • creator_credibility_score: 创作者可信度异常评分
      • content_authenticity_score: 内容真实性异常评分
      • follower_correlation_score: 粉丝相关性异常评分
      • racing_mechanism_score: 赛马机制异常评分
      • fan_growth_score: 粉丝增长异常评分

  • traffic_pool: 流量池分析(TikTok赛马机制)

    • current_tier: 当前流量池级别(1-8)，越高代表流量越大
    • current_tier_name: 当前流量池名称
    • expected_tier: 预期流量池级别，基于有机流量预测
    • expected_tier_name: 预期流量池名称
    • current_views_range: 当前流量池预期观看范围
    • expected_views_range: 预期流量池观看范围
    • estimated_organic_views: 估计有机观看量，扣除虚假后的真实观看

  • suspicious_features: 可疑特征列表，检测到的具体异常现象

  • recommendations: 建议操作

    • action: 建议操作类型，可能值包括:
      • no_action: 无需操作，健康内容
      • monitor: 持续监控，存在轻微异常
      • scheduled_review: 安排审核，存在值得关注的异常
      • immediate_review: 立即审核，存在严重异常
    • risk_level: 风险等级("low", "medium", "high", "critical")
    • potential_revenue_impact: 潜在收益影响
    • suggested_steps: 建议步骤，具体操作建议

  • mcn_report: (可选)MCN商业影响分析报告，适用于商业账号

    • summary: 摘要信息
    • business_impact: 商业影响评估
      • revenue_impact: 收益影响评估
      • brand_safety_impact: 品牌安全影响
      • platform_relationship: 平台关系影响
      • contract_impact: 合约影响评估
    • recommended_actions: 建议操作清单
    • historical_context: 历史背景数据

特性与优势:

  • 基于TikTok原生流量池(Traffic Pool)理论构建的精确评估系统
  • 8个维度、20+指标的全面分析，覆盖流量、互动、创作者、内容等全方位评估
  • 自适应算法，根据账号规模、认证状态、内容类型自动调整阈值
  • 基于大数据统计模型的异常检测，准确识别不自然流量模式
  • 为不同规模账号(微型、小型、中型、大型、超大型)提供定制化评估标准
  • 提供详细的商业影响分析和具体可行的建议步骤

示例响应:

{
  "code": 200,
  "router": "/api/v1/tiktok/analytics/detect_fake_views",
  "params": {
    "item_id": "7460937381265411370",
    "content_category": "verified_large"
  },
  "data": {
    "video_metrics": {
      "total_views": 159414915,
      "total_likes": 15817234,
      "total_comments": 392493,
      "total_favorites": 1051470,
      "total_shares": 1312741,
      "engagement_rates": {
        "like_ratio": 0.09922,
        "comment_ratio": 0.00246,
        "favorite_ratio": 0.0066,
        "share_ratio": 0.00823
      }
    },
    "creator_metrics": {
      "account_age_days": 3733.94,
      "follower_count": 89827771,
      "verified": true,
      "trust_score": 100
    },
    "content_metrics": {
      "content_type": "video",
      "created_by_ai": false,
      "high_quality_upload": true
    },
    "fake_view_analysis": {
      "fake_score": 7.16,
      "confidence_level": "Minimal",
      "estimated_fake_views": 7970745,
      "fake_view_percentage": 5.0,
      "is_suspicious": false,
      "main_detection_reason": "Statistical View Anomalies",
      "component_scores": {
        "engagement_score": 0.0,
        "distribution_score": 10.0,
        "consistency_score": 0,
        "creator_credibility_score": 0,
        "content_authenticity_score": 34.0,
        "follower_correlation_score": 35.0,
        "racing_mechanism_score": 0,
        "fan_growth_score": 45
      }
    },
    "traffic_pool": {
      "current_tier": 8,
      "current_tier_name": "8th-Level Traffic Pool",
      "expected_tier": 8,
      "expected_tier_name": "8th-Level Traffic Pool",
      "current_views_range": "30M+",
      "expected_views_range": "30M+",
      "estimated_organic_views": 148000807
    },
    "suspicious_features": [
      "Suspicious: Reached 100000 followers from 10000 in only 31 days",
      "Suspicious: Account gaining 24063 followers per day on average"
    ],
    "recommendations": {
      "action": "no_action",
      "risk_level": "low",
      "potential_revenue_impact": "minimal",
      "suggested_steps": [
        "No immediate action required",
        "Include in routine monitoring"
      ]
    },
    "mcn_report": {
      "summary": {
        "estimated_revenue_impact": 7970.745,
        "recommended_actions": "No immediate action required"
      },
      "business_impact": {
        "revenue_impact": {
          "level": "low",
          "estimated_amount": 7970.745
        },
        "brand_safety_impact": {
          "level": "minimal"
        },
        "platform_relationship": {
          "status": "good"
        }
      }
    }
  }
}

[English]

Purpose:

  • Analyze TikTok video traffic data using advanced algorithms to precisely detect potential fake views and unnatural engagement
  • Evaluate content authenticity and traffic quality based on TikTok's Traffic Pool theory
  • Provide comprehensive fraud risk analysis with in-depth assessment across 8 dimensions and 20+ metrics
  • Deliver professional traffic quality reports and optimization recommendations for creators, MCN agencies, and content managers

Parameters:

  • item_id: Video ID, required parameter, can be extracted from video URL (e.g., 7460937381265411370 from https://www.tiktok.com/@tiktok/video/7460937381265411370)
  • content_category: Content category, optional parameter, affects engagement rate benchmarks, options include:
    • default: Default category for general content
    • entertainment: Entertainment content, expected to have higher engagement
    • education: Educational content, expected to have moderate engagement and higher save rates
    • product: Product content, expected to have lower engagement but higher conversion
    • verified_large: Large verified accounts, expected to have appropriately lower engagement rates

Return Description:

  • video_metrics: Core video metrics

    • total_views: Total number of views
    • total_likes: Total number of likes
    • total_comments: Total number of comments
    • total_favorites: Total number of saves
    • total_shares: Total number of shares
    • engagement_rates: Engagement rate metrics, higher is better
      • like_ratio: Like rate, normal range 1-10%, may be lower for large accounts
      • comment_ratio: Comment rate, normal range 0.1-0.5%, excellent if above 1%
      • favorite_ratio: Save rate, normal range 0.05-0.8%
      • share_ratio: Share rate, normal range 0.05-0.5%, excellent if above 1%

  • creator_metrics: Creator account health indicators

    • account_age_days: Account age in days, longer is more credible
    • follower_count: Number of followers, affects expected view count
    • verified: Whether account is verified, verified accounts have higher credibility
    • trust_score: Account trust score (0-100), higher is more trustworthy

  • content_metrics: Content quality indicators

    • content_type: Content type (video, image, etc.)
    • created_by_ai: Whether AI-generated, AI-generated content may have specific traffic patterns
    • high_quality_upload: Whether high-quality upload, high-quality uploads are more credible

  • fake_view_analysis: Comprehensive fake traffic analysis

    • fake_score: Fake view score (0-100), lower is better:
      • 0-20: Very low risk, natural traffic patterns
      • 20-40: Low risk, may have minor anomalies but not problematic
      • 40-60: Medium risk, anomalies worth attention
      • 60-80: High risk, obvious fake traffic characteristics
      • 80-100: Very high risk, almost certainly fake traffic
    • confidence_level: Risk level, categorized as "Minimal", "Low", "Medium", "High"
    • estimated_fake_views: Estimated fake views, calculated based on fake traffic model
    • fake_view_percentage: Fake view percentage, proportion of fake views to total views
    • is_suspicious: Whether suspicious, comprehensive judgment if attention is needed
    • main_detection_reason: Main detection reason, most significant anomaly feature
    • component_scores: Dimensional anomaly scores, each 0-100, lower is better:
      • engagement_score: Engagement anomaly score
      • distribution_score: Distribution anomaly score
      • consistency_score: Consistency anomaly score
      • creator_credibility_score: Creator credibility anomaly score
      • content_authenticity_score: Content authenticity anomaly score
      • follower_correlation_score: Follower correlation anomaly score
      • racing_mechanism_score: Racing mechanism anomaly score
      • fan_growth_score: Fan growth anomaly score

  • traffic_pool: Traffic pool analysis (TikTok racing mechanism)

    • current_tier: Current traffic pool level (1-8), higher means more traffic
    • current_tier_name: Current traffic pool name
    • expected_tier: Expected traffic pool level, based on organic traffic prediction
    • expected_tier_name: Expected traffic pool name
    • current_views_range: Current traffic pool expected view range
    • expected_views_range: Expected traffic pool view range
    • estimated_organic_views: Estimated organic views, real views after deducting fake ones

  • suspicious_features: List of suspicious features, specific detected anomalies

  • recommendations: Recommended actions

    • action: Recommended action type, possible values include:
      • no_action: No action needed, healthy content
      • monitor: Continuous monitoring, minor anomalies present
      • scheduled_review: Schedule review, anomalies worth attention
      • immediate_review: Immediate review, serious anomalies present
    • risk_level: Risk level ("low", "medium", "high", "critical")
    • potential_revenue_impact: Potential revenue impact
    • suggested_steps: Suggested steps, specific action recommendations

  • mcn_report: (Optional) MCN business impact analysis report, applicable for business accounts

    • summary: Summary information
    • business_impact: Business impact assessment
      • revenue_impact: Revenue impact assessment
      • brand_safety_impact: Brand safety impact
      • platform_relationship: Platform relationship impact
      • contract_impact: Contract impact assessment
    • recommended_actions: Recommended action list
    • historical_context: Historical background data

Features and Advantages:

  • Precise evaluation system built on TikTok's native Traffic Pool theory
  • Comprehensive analysis across 8 dimensions and 20+ metrics, covering traffic, engagement, creator, content, etc.
  • Adaptive algorithm automatically adjusts thresholds based on account size, verification status, content type
  • Anomaly detection based on big data statistical models, accurately identifies unnatural traffic patterns
  • Provides customized evaluation standards for different account sizes (micro, small, medium, large, extra-large)
  • Delivers detailed business impact analysis and specific, actionable recommendations

Example Response:

{
  "code": 200,
  "router": "/api/v1/tiktok/analytics/detect_fake_views",
  "params": {
    "item_id": "7460937381265411370",
    "content_category": "verified_large"
  },
  "data": {
    "video_metrics": {
      "total_views": 159414915,
      "total_likes": 15817234,
      "total_comments": 392493,
      "total_favorites": 1051470,
      "total_shares": 1312741,
      "engagement_rates": {
        "like_ratio": 0.09922,
        "comment_ratio": 0.00246,
        "favorite_ratio": 0.0066,
        "share_ratio": 0.00823
      }
    },
    "creator_metrics": {
      "account_age_days": 3733.94,
      "follower_count": 89827771,
      "verified": true,
      "trust_score": 100
    },
    "content_metrics": {
      "content_type": "video",
      "created_by_ai": false,
      "high_quality_upload": true
    },
    "fake_view_analysis": {
      "fake_score": 7.16,
      "confidence_level": "Minimal",
      "estimated_fake_views": 7970745,
      "fake_view_percentage": 5.0,
      "is_suspicious": false,
      "main_detection_reason": "Statistical View Anomalies",
      "component_scores": {
        "engagement_score": 0.0,
        "distribution_score": 10.0,
        "consistency_score": 0,
        "creator_credibility_score": 0,
        "content_authenticity_score": 34.0,
        "follower_correlation_score": 35.0,
        "racing_mechanism_score": 0,
        "fan_growth_score": 45
      }
    },
    "traffic_pool": {
      "current_tier": 8,
      "current_tier_name": "8th-Level Traffic Pool",
      "expected_tier": 8,
      "expected_tier_name": "8th-Level Traffic Pool",
      "current_views_range": "30M+",
      "expected_views_range": "30M+",
      "estimated_organic_views": 148000807
    },
    "suspicious_features": [
      "Suspicious: Reached 100000 followers from 10000 in only 31 days",
      "Suspicious: Account gaining 24063 followers per day on average"
    ],
    "recommendations": {
      "action": "no_action",
      "risk_level": "low",
      "potential_revenue_impact": "minimal",
      "suggested_steps": [
        "No immediate action required",
        "Include in routine monitoring"
      ]
    },
    "mcn_report": {
      "summary": {
        "estimated_revenue_impact": 7970.745,
        "recommended_actions": "No immediate action required"
      },
      "business_impact": {
        "revenue_impact": {
          "level": "low",
          "estimated_amount": 7970.745
        },
        "brand_safety_impact": {
          "level": "minimal"
        },
        "platform_relationship": {
          "status": "good"
        }
      }
    }
  }
}

Imported from a validated TikHub OpenAPI document. Pricing must be reviewed before enabling.

Try 检测视频虚假流量分析/Detect fake views in video

Test this model in the Sandbase Playground with your own prompts.

Open in Playground

Related Models

获取视频评论关键词分析/Get comment keywords analysisTikTokapi获取创作者信息和里程碑数据/Get creator info and milestonesTikTokapi获取作品的统计数据/Get video metricsTikTokapi获取单个广告详情/Get single ad detailTikTokapi获取搜索板块配置/Get search modules configTikTokapi获取支持的国家地区列表/Get supported location listTikTokapi获取查询建议/Get query suggestionsTikTokapi获取推荐广告/Get recommended adsTikTokapi