We did some machine learning with standard random forest classifier on the study genes and then on genes specific to relationships we saw in the data analysis of groups and subgroups of genes. Not all subsets were done, but some were terrible and some just not good at all. The best performer was the genes from the study. The avg/median of groups and subgroups didn't seem to work very well. The best score was 70% with genes of study and 60% with some genes that seemed to be more associated with males. It seems that the 4 class model on the proposed 4 classes scored 30% with the genes in the study and 40% in the male gene predictor in the male inclusive samples. The transition state seems likely but at times it misidentified an mDLBCL class as it for IFNG-L when it should be CHL or pDLBCL. It's apparent that the large B-cell lymphomas are EBV associated and so is CHL but they remain difficult to classify even with a classifier tool.

Tugas Mata Kuliah Sistem Informasi Manajemen

Justitsar Bella Dina - M0725068 - Kelas E

Este documento presenta un procedimiento completo para la simulación y estimación del modelo Tucker3 en R. Se genera un arreglo de tres vías a partir de matrices de componentes y un núcleo, incorporando un término de error. Luego, se construyen las matricizaciones correspondientes a cada modo y se implementa un algoritmo iterativo basado en descomposición en valores singulares para estimar modelos candidatos con distintas dimensiones. La selección del modelo se realiza mediante el scree plot multiway, evaluando el equilibrio entre ajuste y parsimonia. Finalmente, se calculan medidas de ajuste y contribuciones por componente, permitiendo comparar la estructura estimada con la verdadera utilizada en la simulación.

Laporan ini menyajikan analisis data pada Coffee Chain Dataset dengan fokus pada hubungan antara Sales (penjualan) dan Profit (keuntungan), serta performa produk berdasarkan Product Line dan Product Type. Analisis dilakukan melalui beberapa tahapan, yaitu eksplorasi data, analisis statistik deskriptif, uji korelasi, serta visualisasi data menggunakan scatter plot, boxplot, histogram, dan bar plot. Hasil analisis menunjukkan bahwa terdapat hubungan positif antara Sales dan Profit, di mana peningkatan penjualan cenderung diikuti oleh peningkatan keuntungan. Selain itu, produk dalam kategori Beans memiliki performa profit yang lebih tinggi dibandingkan Leaves, serta terdapat perbedaan kontribusi penjualan antar Product Type. Temuan ini diharapkan dapat memberikan wawasan dalam memahami pola penjualan dan membantu pengambilan keputusan bisnis yang lebih efektif.