本研究旨在探討結合人格特質與海報主色於類神經網路推薦電影之效果。我們使用了Moivelens資料庫的用戶大五人格特質資料，從IMDb網站上蒐集電影海報資料，透過影像處理技術提取出海報主要色彩，再透過多層感知器回歸分析，以推薦符合用戶人格特質與喜好的電影。 實驗結果顯示，將電影評分、人格特質和海報主色結合後，推薦系統的準確性有所提升。回歸分析的結果顯示，此三種資料的解釋力達34.2%，RMSE為0.816，MSE為0.666。 綜合以上結果，本研究提出了一個結合電影評分、人格特質和海報主色的推薦系統，可透過多層感知器回歸分析提供更精確的推薦服務。此推薦系統可為電影愛好者提供更加個性化和符合自身喜好的電影推薦。 ;This study aims to explore the effectiveness of combining personality traits and poster dominant colors in neural network-based movie recommendations. We used the Big Five personality traits test data from the MovieLens database and collected movie poster data from the IMDb website. Through image processing techniques, we extracted the dominant colors of the posters and used a multi-layer perceptron regression analysis to recommend movies that match user′s personality traits and preferences. The experimental results showed that the accuracy of the recommendation system improved when combining movie ratings, personality traits, and poster dominant colors. The regression analysis results showed that the explanatory power of these three types of data reached 34.2%, with an RMSE of 0.816 and MSE of 0.666. Based on these results, this study proposes a recommendation system that combines movie ratings, personality traits, and poster dominant colors, which can provide more accurate recommendation services through multi-layer perceptron regression analysis. This recommendation system can provide movie enthusiasts with more personalized and tailored movie recommendations that align with their preferences.
