智慧型手機的誕生帶來人們生活上的便利，卻也成為詐欺犯罪新興的溫床，詐欺犯罪往往侵害人民財產甚巨，因為詐欺犯罪會隨著時間及科技進步，逐漸演進成不同型態的詐欺犯罪，從傳統詐欺犯罪、電信詐欺犯罪、網路詐欺犯罪，演進到利用智慧型手機當成標的物的智慧型手機犯罪，也因為智慧型手機結合了通信及網路功能，所以其成為目前最新興的詐欺犯罪型態，如何防制智慧型手機上的詐欺犯罪，己是刻不容緩。

詐欺犯罪不論如何演進，利用何種工具輔助犯罪，其均需要透過媒介進行，將其所欲達成的目的轉化為訊息傳遞予犯罪目標，所以詐欺犯罪最主要的關鍵在於訊息的傳遞及互動，當詐欺訊息無法傳遞予被害人，或是訊息在傳遞時能事先予接收者預警，這樣定能有效阻止詐欺犯罪的進行，而詐欺訊息的組成其關鍵字包含特定名詞、特定動詞及網址連結，因為詐欺犯罪者往往將重覆的訊息大量傳遞，故其所傳遞的詐欺訊息內容都不外乎上述三類關鍵字所組成的訊息。

而目前智慧型手上的詐欺訊息大致分為點數卡詐欺、小額付款詐欺及惡意連結詐欺等三類，本研究擬藉由關鍵字擷取，基於目前市面上較普遍使用的Android智慧型手機，進行詐欺犯罪的探討，試圖透過訊息關鍵字的擷取統計與歸納，找出詐欺訊息進行預警，並輔以關鍵字權重判斷進行實驗，利用層級分析法針對詐欺訊息三項指標特徵進行權重配分，以提高詐欺訊息的辨別率，透過發現詐欺訊息來警示繼而減少民眾誤信詐欺訊息內容，針對智慧型手機詐欺提出有效的防制措施，有效降低智慧型手機詐欺犯罪的發生率。
The invention of smart phone brings the convenience of human life; however, it becomes a new approach for fraud which it always leads to the great loss of properties. Due to new technology, different types of fraud, from traditional ones, telecommunication, the Internet to smart phone, are gradually developed and changed with time. Fraud through smart phone is the newest approach in present time because it combines telecommunication with the Internet functions. Hence, it is urgent to figure out the solution how to prevent frauds through smart phone.
No matter how frauds becomes, what kinds of tools used, it has to via a media to convert the main purpose into message and deliver to the crime target. Therefore, the main focus on fraud is transmission and interaction of message. Frauds can be effective deterred when the message fails to be transmitted or alarmed by the receiver in advance. Specific nouns, verbs and the URL links are the three major keywords consisting a fraudulent message because the scammer tends to deliver huge amount of repetitive remarks.
Presently, the approach of fraud through smart phone can be divided into three major categories: Game card fraud, small payments fraud, and malware link fraud. This research aims to investigate the fraud by keywords extraction on the common-used Android system. Moreover, the analytic hierarchy process (AHP) would be conducted to weight and get a score by statistics and induction of these keywords in order to make fraudulent messages identifiable. The final purpose is to reduce the frequency of fraud on smart phone effectively by assisting to identify the fraudulent messages.
資料來源https://ndltd.ncl.edu.tw/cgi-bin/gs32/gsweb.cgi/ccd=Orl8Fo/record?r1=1&h1=22