本文修正Collaborative Click Fraud Detection and Prevention system (CCFDP)的計分方式，除了原本無效點擊的證據外，加入有效點擊的證據，並對模型進行測試。測試內容包含模擬手動點擊與透過不同IP的機器人點擊(bot clicks)，以及CCFDP系統在不同數目的時間分段下，手動點擊與機器人點擊點擊次數對分數的關係。結果顯示，點擊分數受到不同時間分段的影響，CCFDP在適當的時間分段下，能判斷出手動點擊與機器人點擊。
The scoring method of a Collaborative Click Fraud Detection and Prevention sys-tem (CCFDP) was modified in this research, and evidence of effective clicks was added into and used to test the model in addition to the original evidence on ineffective clicks. The test content included simulated manual clicks and bot clicks via different IPs, as well as the effects of a number of manual clicks and a number of bot clicks on scores when using the CCFDP system under different segmentation methods and different numbers of time segments. The results reveal that the click scores are affected by dif-ferent segmentation methods and different numbers of time segments and that CCFDP can differentiate between a manual click and a bot click under appropriate time seg-mentation.
