Keen Learner and Exp... • 2m
Day 7 of learning AI/ML as a beginner. Topic: One Hot Encoding and Future roadmap. Now that I have learnt how to clean up the text input a little its time for converting that data into vectors (I am so glad that I have learned it despite getting criticism on my approach). There are various processes to convert this data into useful vectors: 1. One hot encoding 2. Bag of words (BOW) 3. TF - IDF 4. Word2vec 5. AvgWord2vec These are some of the ways we can do so. Today lets talk about One hot encoding. This process is pretty much outdated and is rarely used in real word scenarios however it is important to know why we don't use this and why are there different ways? One hot encoding is a technique used for converting a variable into a binary vector. Its advantage is that it is easy to use in python via scitkit learn and pandas library. Its disadvantages however includes. sparse matrix which can lead to overfitting(when a model performs well on the data its been trained and performs poorly with new one). Then it require only fixed sized input in order to get trained. One hot encoding does not capture sematic meaning. And what about a word being out of the vocabulary. Then it is also not practical to use in real world scenarios as it is not much scalable and may lead to problems in future. I have also attached my notes here explaining all these in much details.




I'm just a normal gu... • 6m
Fintech unicorn Razorpay has taken a significant step toward its initial public offering (IPO) by converting into a public company. According to a regulatory filing, the company secured approval from its members during an extraordinary general meeti
See More
Hey I am on Medial • 1y
Thanks @ZeptoNow for its new feature Zepto Cafe making it easy to deliver food within 10min, and the food is meant to be consumed when it's hot. It would be nice if you add more food options like Dosa, Biriyani, and more. #zepto #zeptocafe #Instantfo
See More

Download the medial app to read full posts, comements and news.