Essential guide to LSAN: Modeling Long-term dependencies and Short-term correlations with Hierarchical Attention — Electronic Health Records (EHR) are comprehensive historical health records that contain the symptoms of a patient when he/she visits a doctor. EHR data has a two-level hierarchical structure that consists of a set of time-ordered visits, and within each visit, there is a set of unordered diagnosis codes. …

Essential guide to time series analysis — The real-world data often contain missing values. All types of the dataset including time-series data have the problem with missing values. The cause of missing values can be data corruption or failure to record data at any given time. In one of my previous articles, I have discussed 7 different…

Text data augmentation for NLP tasks — Training on a small sample of data increases the chances of overfitting. Data augmentation is a technique to create artificial similar samples of existing data. Data augmentation techniques are often used for tasks where the model expects a large amount of data, but we have limited access to the data…

Scalable hyperparameter optimization framework that solves the pain to optimize the ANN model — To tune the machine learning algorithms, the Scikit-learn package offers the implementation of GridSearchCV, RandomSearchCV to find the optimal set of hyperparameters for ML models. But when it comes to training a deep learning (ANN) model, the data scientist often struggles to find the right number of hidden layers, number…

Essential guide to perform NER using pre-trained or custom trained model — NLP tasks that involve an automatic understanding of the text, need Named Entity Recognition (NER) in its pipeline. NER models locate and classify named entities in a given text corpus. A named entity can be a person, organization, location, language, etc. NER models can be used to understand the meaning…

Essential guide to imputation strategies — A real-world dataset often contains a lot of missing values that may be caused due to data corruption or failure to records data. The presence of missing values in the data prevents training a robust machine learning model. …

Scikit-learn algorithm cheat-sheet — Model training is an important component of a data science model development pipeline. Choosing the right machine learning algorithm can be a tedious and confusing task for a data scientist. There is no such solution or one approach to pick an estimator that fits all kinds of data. Choosing the…

No need to handle missing values explicitly — A real-world dataset often contains a lot of missing values that may be caused due data corruption or failure to record data. The presence of missing values in the data may impact the robustness of the trained model. A data scientist needs to handle the missing values explicitly during the…

Data Augmentation in NLP with NLPAUG — Image augmentation is a common task performed on the raw image data prior to training a CNN model. Image augmentation refers to manipulating the images including rotation, shearing, flipping, gray scaling, etc. Due to the semantically invariant transformation, the performance of the CNN models improves, and hence image augmentation is…