sql_basic-queries_

MySQL

survival_analysis_from_scratch

An introduction to the concepts of Survival Analysis and its implementation in lifelines package for Python.

svm_hyperparameter_tuning_kernel_tricks

Effect of Gamma values and C values visualization on dataset and errors/misclassification

t-sne_and_umap_visuals

t-SNE (pronounced tiz-knee), which stands for t-distributed Stochastic Neighbor Embedding was proposed much more recently by Laurens van der Maaten and Geoffrey Hinton in their 2008 paper. This works in a similar way to PCA but has some key differences: Firstly, this is a stochastic method. So if you run multiple t-SNE plots on the same dataset it can look different. Another difference is that this is an iterative method. It works by repeatedly moving datapoints closer or further away from each other depending on how 'similar' they are. The new representation is non-linear. This makes it harder to interpret but it can be very effective at 'unravelling' highly non-linear data. The main downside to t-SNE is that is very slow compared to the other dimensionality techniques. This is because it makes calculations on a pair-wise basis, which does not scale well with large datasets.

text_mining_nlp

Natural Language Processing

time_series_forecasting_from_scratch

web_scraping_amazon_product_reviews

Web Scraping Amazon Reviews with Multiple Pages loop till the Last Page

web_scraping_amazon_products_image_and_url

E-commerce companies use recommendation systems to provide suggestions to the customers. They use item-item collaborative filtering, which scales to massive datasets and produces high quality recommendation systems in the real time. This system is a kind of an information filtering system which seeks to predict the "rating" or preferences which user is interested in.