MY BLOG

The cancer researcher who is an avid lover of data science and AI applications

A little about myself, Feb 29, 2020

My background is in cancer research. I obtained both my Bachelor and PhD. of Science from Nanyang Technological University (Singapore). I work on multiple aspects of cancer research. With the vast amount of complex data generated from genome-wide and high throughput cell imaging , I have also realized the need for using Artificial Intelligence to streamline the analysis of phenotypic and NGS data. In order to address that aspect of my current work to improve on my work efficiency, I have undertaken training in AI and python programming by taking a part-time Specialist Diploma in Data Science (Artificial Intelligence) from Singapore Polytechnic. In my spare time after work and course, I have also self-directed my learning to learn other programming languages such as R to incorporate what I have learned with my current cancer NGS research work (R is the preferred programming language for bioinformatics analysis). I have self-taught myself on how to analyze the high-dimensional and complicated NSG datasets and have deployed some machine learning algorithms to retrieve data of interest to gain some novel insights on my work. My goal is to incorporate AI applications with cancer research to gain novel data-driven insights (biomarkers of drug resistance, etc.) to improve the high failure rate of drug discovery problems.

Photo by Franki Chamaki on Unsplash

My First Blog Entry

UNSUPERVISED LEARNING of the IRIS dataset with Kmeans Clustering Methods¶ Feb 29, 2020

In this post, I will describe how to perform an unsupervised learning experiment on the Iris dataset. The Problem: Given the Iris dataset, if we knew that there were k types of iris, but did not have a taxonomist colleague to label them: we could attempt an unsupervised clustering task: split the observations into well-separated group called clusters. We will use the Iris dataset to make a prediction of the number of different. The dataset contains a set of 150 records with four features —  petal length, petal width, sepal length, sepal width, and three iris classes: setosa, virginica and versicolor. We'll feed the four features of our flower to the unsupervised algorithm and it will predict which class the iris belongs.

My Second Blog Entry

UNSUPERVISED LEARNING of the IRIS dataset by Hierachical Clustering¶ March 1, 2020

In this post, I will describe how to perform an unsupervised learning experiment on the Iris dataset by Hierachical Clustering.

My Third Blog Entry

UNSUPERVISED LEARNING of the IRIS dataset by Gaussian Mixture Model¶ March 2, 2020

In this post, I will describe how to perform an unsupervised learning experiment on the Iris dataset by Gaussian Mixture Model.

My Fourth Blog Entry

MNIST Handwritten digit classification problem by Deep Learning March 24, 2020

In this post, I will describe how to perform Convolutional Neural Network on the MNIST Handwritten digit classification problem.