在現代社會中，許多家庭都有飼養寵物，使得何時該餵食寵物的問題也成為諸多飼主的一大疑問。不僅是家中寵物的飼主，對於動物園中的飼養員或是動物收容所的養育員皆是一大難題。倘若我們能明確地知道動物叫聲的意義，進而判別動物是否正處於飢餓狀態，便可在適當的時間給予動物餵食。由於人類無法正確明白動物的叫聲，因此我們選擇使用機器學習的方式來判別動物是否處於飢餓狀態。
我們使用的資料集取自於取自POND5影音資料庫，此資料庫收錄數以千計的多媒體檔案讓使用者下載，而這些素材最大的特色是不受版權限制，更吸引使用者的特點是，此資料庫內建搜尋與篩選器的功能，讓使用者可以透過篩選每個多媒體檔案上帶有的標籤過濾出想使用的多媒體檔案。所以我們透過資料庫已貼上的標籤，將資料集分為兩組，分別是飢餓（hungry）組與非飢餓（not hungry）組，且兩組資料中亦帶有此資料庫預先判別之情緒標籤，如飢餓生氣的狼等，再讓機器根據情緒變化判斷各個音訊檔分別屬於何種狀態，最後經由比對機器分類的結果與原始檔案帶有的標籤輸出分類的正確比率。
此篇論文的主要貢獻在於我們分析出動物叫聲中的情緒變化，讓分類器經由情緒變化判斷此聲音是否為飢餓狀態，而輸入資料則含有飢餓與非飢餓狀態的標籤。在第一層分類器中，我們讓利用它幫助我們做情緒辨識與記錄情緒變化，進而輸出情緒變化矩陣，因此我們稱其為情緒分類器。此外，我們根據動物相關文獻得知負面的情緒會成動物的飢餓狀態，因此在最後的分類器中，我們讓分類器依照負面情緒與否判定是否為飢餓狀態，並使用支援向量機（SVM）、決策樹（Decision Tree）與隨機森林（Random Forest）作為分類的方法。

In modern society, many families have pets, so the question of when to feed pets has become a big question for many owners. Not only is the owner of the pets in the house, but it is also a big problem for the breeders in the zoo or the breeders in the animal shelters. If we can know the meaning of the animal's cry clearly, and then determine whether the animal is starving, we can feed the animal at an appropriate time. Since humans cannot understand the cry of animals correctly, we choose to use machine learning to determine whether animals are hungry.
The data set we use is taken from the POND5 audio-visual database. This database contains thousands of multimedia files for users to download. The biggest feature of these materials is that they are not restricted by copyright and are more attractive to users. , This database has built-in search and filter functions, allowing users to filter out the multimedia files they want to use by filtering the tags on each multimedia file. Therefore, we divide the data set into two groups based on the tags that have been attached to the database, namely the hungry group and the not hungry group, and the two groups of data also contain the emotions pre-identified by the database. Labels, such as hungry wolves, let the machine determine which state each audio file belongs to based on emotional changes, and finally output the correct ratio of classification by comparing the classification result of the machine with the label with the original file.
The main contribution of this paper is that we analyze the emotional changes in animal calls, and let the classifier determine whether the voice is in a hungry state through the emotional changes, and the input data contains the labels of the hungry and non-hungry states. In the first-level classifier, we use it to help us identify emotions and record emotion changes, and then output the emotion change matrix, so we call it an emotion classifier. In addition, we know from animal-related literature that negative emotions will become animal hunger states, so in the final classifier, we let the classifier determine whether it is a hungry state according to whether negative emotions are or not, and use a support vector machine (SVM), Decision Tree (Decision Tree) and Random Forest (Random Forest) as classification methods.

壹、 緒論……………………………………………………………..………………..1
1.1 動機與目的………………………………………………………………...1
1.2 論文架構…………………………………………………………………...2
貳、 相關研究探討………………………………………..…………………………..3
2.1 人工智慧…………………………………………………………………...3
2.2 機器學習…………………………………………………………………...3
2.3 深度學習…………………………………………………………………...4
參、 系統架構……………………………………………..…………………..............7
3.1 整體架構…………………………………………………………...………7
3.2 情緒分類器……………………………………………….……………..…9
3.3 決策分類器……………………………………….………………....……12
3.3.1 支援向量機………………………………….…………..…...……12
3.3.2 決策樹…………………………………………………………..…15
3.3.3 隨機森林…………………………………………………………..17
肆、 實驗與結果…………………………………………………………..…………21
4.1 資料集…………………………………………………………………….21
4.2 實驗方法………………………………………………………………….22
4.3 實驗結果………………………………………………………………….22
伍、 結論……………………………………………………………………………..24
參考文獻……………………………………………………………………………..25

[1] Teresa Bradley Bays, Teresa Lightfoot, Jörg Mayer, “Exotic Pet Behavior: Birds, Reptiles, and Small Mammals”, 2006.
[2] Pierluca Costa, Elisabetta Macchi, Laura Tomassone, Fulvio Ricceri, Enrico Bollo, Frine Eleonora Scaglione, Martina Tarantola, Michele De Marco, Liviana Prola, Domenico Bergero, Achille Schiavone, “Feather picking in pet parrots: sensitive species, risk factor and ethological evidence”, 2016.
[3] Aaron Katcher, Alan Beck, “Between Pets and People: The Importance of Animal Companionship”, 1996.
[4] Sharon Crowell-Davis, “Motivation for Pet Ownership and Its Relevance to Behavior Problems”, 2008.
[5] Ephrat Livni, “What the study of animal emotions shows us about being human”, 2019.
[6] Temple Grandin, Catherine Johnson, “Animals Make Us Human: Creating the Best Life for Animals”, 2009.
[7] Jennifer Kay MacCormack, “Feeling hangry? When hunger is conceptualized as emotion.”, 2019.
[8] Sanjay Awathale, Biru Dudhbhate, Rakesh Rahangdale, Chandrashekhar Borkar, Nishikant Subhedar, Dadasaheb Kokare, “Denial of food to the hungry rat: A novel paradigm for induction and evaluation of anger-like emotion”, 2020.
[9] Franklin McMillan, “Stress-induced and emotional eating in animals: A review of the experimental evidence and implications for companion animal obesity”, 2013.
[10] Ying-Li Tian, Takeo Kanade, “Facial Expression Analysis. In: Handbook of Face Recognition”, 2001.
[11] Ossama Abdel-Hamid, Abdel-rahman Mohamed, Hui Jiang, Li Deng, Gerald Penn, Dong Yu, “Convolutional Neural Networks for Speech Recognition”, 2014.
[12] Chu-hsuan Lin, Yen-Sheng Chen, “Speech Emotion Recognition via Nonlinear Dynamical Features”, 2015.
[13] Veronica Pérez¬Rosas, Mohamed Abouelenien, Rada Mihalcea, Mihai Burzo, “Deception detection using real¬life trial data”, 2015.
[14] Mandar Gogate, Ahsan Adeel, Amir Hussain, “Deep learning driven multimodal fusion for automated deception detection”, 2017.
[15] Pariwat Ongsulee, “Artificial intelligence, machine learning and deep learning”, 2017.
[16] Julian Togelius, Jesper Juul, Geoffrey Long, William Uricchio, Mia Consalvo, “3 What Is (Artificial) Intelligence?”, 2018.
[17] Susmita Ray, “A Quick Review of Machine Learning Algorithms”, 2019.
[18] Foram Shah, Vibha Patel, “A review on feature selection and feature extraction for text classification”, 2016.
[19] Jingyang Gao, Qunxiong Zhu, Chenglizhao Chen, “A new method of evaluation and optimization for feature extraction”, 2010.
[20] Michael Fu, “AlphaGo and Monte Carlo tree search: The simulation optimization perspective”, 2016.
[21] Guosheng Yang, Yingzi Lin, Chunli Ji, Madan Gupta, “On Definition of Deep Learning”, 2018.
[22] Guang-Xin Lou, Hong-Zhen Shi, “Face image recognition based on convolutional neural network”, 2020.
[23] Jia-Huan Zhou, Di Xiao, Meng-Yi Zhang, “Feature Correlation Loss in Convolutional Neural Networks for Image Classification”, 2019.
[24] Nipa Chowdhury, Mohammod Abul kashem, “A comparative analysis of Feed-forward neural network & Recurrent Neural network to detect intrusion”, 2008.
[25] Haruhiko Takase, “Time sequential pattern transformation and attractors of recurrent neural networks”, 1993.
[26] Shikha Gupta, Jafreezal Jaafar, Wan Fatimah wan Ahmad, Arpit Bansal, “Feature Extraction using MFCC”, 2013.
[27] Carles Bond, Bella DePaulo, “Accuracy of deception judgments”, 2006.
[28] Sonali Wankhede, “Analytical Study of Neural Network Techniques: SOM, MLP and Classifier-A Survey”, 2014.
[29] Miguel Antonio Sovierzoski, Fernanda Isabel Marques Argoud, Fernando Mendes de Azevedo, “Evaluation of ANN Classifiers During Supervised Training with ROC Analysis and Cross Validation”, 2008.
[30] Robert Hecht-Nielsen, “Neural Networks for Perception”, 1992.
[31] Vikramaditya Jakkula, “Tutorial on Support Vector Machine (SVM)”, 2006.
[32] Chong-Yung Chi, Wei-Chiang Li, Chia-Hsiang Lin, “Convex Optimization for Signal Processing and Communications: From Fundamentals to Applications.”, 2017.
[33] David Magerman, “Statistical Decision-Tree Models for Parsing”, 1995.
[34] Luis Belanche, Angela Nebot, “Feature selection algorithms: a survey and experimental evaluation”, 2002.
[35] Samuel Yeom, Irene Giacomelli, Matt Fredrikson, Somesh Jha, “Privacy Risk in Machine Learning: Analyzing the Connection to Overfitting”, 2018.
[36] Vrushali Kulkarni, Pradeep Sinha, “Pruning of Random Forest classifiers: A survey and future directions”, 2012.
[37] Andreas Weingessel, Evgenia Dimitriadou, Kurt Hornik, “An Ensemble Method for Clustering”, 2003.
[38] Bernd Fischer, Joachim Buhmann, “Bagging for path-based clustering”, 2003.
[39] Andreas Mayr, Harald Binder, Olaf Gefeller, Matthias Schmid, “The Evolution of Boosting Algorithms - From Machine Learning to Statistical Modelling”, 2014.
[40] Zeng-Chang Qin, “ROC analysis for predictions made by probabilistic classifiers”, 2005.
[41] Longbiao Chen, Hai-Lin Tang, “Improved computation of beliefs based on confusion matrix for combining multiple classifiers”, 2004.