不斷增加的專利訴訟案件數量顯示近年來科技產業對智慧財產權保護意 識日益高漲,許多公司為保衛自己的權益而戰,專利儼然已成為企業與他人競 賽時重要的策略之一;由許多專利訴訟案件中顯示,專利訴訟發生時皆造成公 司股價之變動,故企業究竟應如何掌握適當時機以回應專利訴訟對公司股價變 動之影響?本研究認為此議題值得深入探討。 本研究的研究目的為探討當公司在發生專利訴訟時,公司股價變動的預測 模式,首先先藉由事件研究法確認專利訴訟與股價的關係,找出專利訴訟事件 發生時公司股價變動的幅度以及影響股價的天數;接著再以倒傳遞類神經網路 建立發生專利訴訟事件時公司股價變動的預測模式。 本研究提供公司未來準備提出專利訴訟或發生專利訴訟時,能掌握事件對 股價影響的時間,由所蒐集之股價資料進行事件研究法分析的結果,當公司發 生專利訴訟事件時,對股價影響最顯著的天數為事件發生後的第一天及第二 天,事件約可影響五天左右,所以不論是原告還是被告公司,都要抓緊專利訴 訟所影響的黃金時間(五天之內),尤其是在威力最強的第一天及第二天,所有 的策略與回應機制都要控制在五天內要能達到效果,否則將錯過市場反應的時 間;同時經由本研究測試訓練後所得到的最適類神經網路模型可以用來預測公 司股價之變動;最後,提供建議給專利訴訟的原告與被告公司如何利用本研究 所得到的影響天數、股價變動幅度以及類神經網路的預測模型進行專利訴訟策 略及回應機制的擬定。
The more and more cases accelerating in the patent litigation could awake the protecting of the intellectual property rights. The enterprises have been fight for the patent litigation game. The way in patent litigation, the company's value have been deeply influenced. We are interested in how to handhold the timing to utilize the announcement day's information to make market value more valuable? This study would deeply discuss this issue. There are two purposes of this study, first is to make sure the relationship between patent litigation and stock prices, then finding the event’s influenced days and variation range. Second is to set a well-forecasting model to forecast the variation range of stock prices when patent litigation happens. The event study and Back-Propagation Network are used in this study. The results of the event study would show the variation range and the influence day of stock when patent litigation happens, and the well-forecasting model that trained by Back-Propagation Network could be utilized in future forecast and feedback control. When patent litigation happens, the company could utilize these two results to evaluate the influence and hold the golden timing to make strategy to fight or defense.