Perbandingan Performa Model Data Mining untuk Prediksi Dropout Mahasiwa

Penentuan teknik/model data mining yang tepat pada sebuah kasus sangat penting untuk mendapatkan model yang baik (tingkat akurat tinggi dan kesesuaiannya dengan masalah yang dipecahkan). Penelitian ini bertujuan untuk membandingkan performa teknik data mining untuk diterapkan pada kasus prediksi dropout mahasiswa. Perbandingan performa dilakukan menggunakan library PyCaret pada Python untuk melakukan pemodelan menggunakan 14 model / teknik data mining yaitu: Extreme Gradient Boosting, Ada Boost Classifier, Light Gradient Boosting Machine, Random Forest Classifier, Gradient Boosting Classifier, Extra Trees Classifier, Decision Tree Classifier, K Neighbors Classifier, Naive Bayes, Ridge Classifier, Linear Discriminant Analysis, Logistic Regression, SVM - Linear Kernel, dan Quadratic Discriminant Analysis. Metrik evaluasi performa model yang digunakan yaitu Accuracy, AUC, Recall, Precision, F1, Kappa, dan MCC (Matthews correlation coefficient). Hasil eksperimen menunjukkan bahwa kasus prediksi dropout mahasiswa lebih tepat jika dimodelkan dengan model berbasis ensemble learner dan pohon keputusan dengan akurasi mencapai 99%. Pohon keputusan memiliki keunggulan dibandingkan model lain seperti SVM - Linear Kernel dan Quadratic Discriminant Analysis karena ia dapat dengan lebih detil dalam memisahkan data ke dalam kedua kelas target. Setelah dilakukan penyesuaian atribut, pembuangan data dengan missing values, dan parameter tuning, didapatkan hasil akurasi yang mirip dari berbagai model yaitu sebesar 87%. Perbedaan akurasi antar model menjadi sangat kecil di saat atribut data yang digunakan sedikit.

On Clustering for Cell Phenotyping in Multiplex Immunohistochemistry (mIHC) and Multiplexed Ion Beam Imaging (MIBI) Data

Problem: Multiplex immunohistochemistry (mIHC) and multiplexed ion beam imaging (MIBI) images are usually phenotyped using a manual thresholding process. The thresholding is prone to biases, especially when examining multiple images with high cellularity. Results: Unsupervised cell phenotyping methods including PhenoGraph, flowMeans, and SamSPECTRAL, primarily used in flow cytometry data, often perform poorly or need elaborate tuning to perform well in the context of mIHC and MIBI data. We show that, instead, semi-supervised cell clustering using Random Forests, linear and quadratic discriminant analysis are superior. We test the performance of the methods on two mIHC datasets from the University of Colorado School of Medicine and a publicly available MIBI dataset. Each dataset contains numerous highly complex images.

A Comparative Study of the Impact of Data Augmentation in Machine Learning Based Classification Accuracy

Traumatic Brain Injury is the primary cause of death and disability all over the world. Monitoring the intracranial pressure (ICP) and classifying it for hypertension signals is of crucial importance. This thesis explores the possibility of a better classification of the ICP signal and detection of hypertensive signal prior to the actual occurrence of the hypertensive episodes. This study differ from other approaches astime series is converted into images by Gramian angular field and Markov transition matrix and augmented with data. Due to unbalanced data, the effect of smote extended nearest neighbour algorithm for balancing the data is examined. We use various machine learning algorithms to classify the ICP signals. The results obtained shoe that Ada boost performance is the best among compared algorithms. F1 score of the Ada boost is 0.95 on original dataset, and 0.9967 on balanced and augmented dataset. Quadratic Discriminant Analysis F1 score is 1 when data is augmented and balanced.

A Comparative Study of the Impact of Data Augmentation in Machine Learning Based Classification Accuracy

Traumatic Brain Injury is the primary cause of death and disability all over the world. Monitoring the intracranial pressure (ICP) and classifying it for hypertension signals is of crucial importance. This thesis explores the possibility of a better classification of the ICP signal and detection of hypertensive signal prior to the actual occurrence of the hypertensive episodes. This study differ from other approaches astime series is converted into images by Gramian angular field and Markov transition matrix and augmented with data. Due to unbalanced data, the effect of smote extended nearest neighbour algorithm for balancing the data is examined. We use various machine learning algorithms to classify the ICP signals. The results obtained shoe that Ada boost performance is the best among compared algorithms. F1 score of the Ada boost is 0.95 on original dataset, and 0.9967 on balanced and augmented dataset. Quadratic Discriminant Analysis F1 score is 1 when data is augmented and balanced.