{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/d85e8dd15837430eb726ad0852451773\" frameborder=\"0\" width=\"1728\" height=\"1296\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":1296,"width":1728,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":1296,"thumbnail_width":1728,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/d85e8dd15837430eb726ad0852451773-db1c1f0cd278f109.gif","duration":213.385,"title":"Potability Predictions: Analyzing Water Safety Based on Chemical Features","description":"Hello, my name is Ema, and in this video, I discuss my project on potability predictions, where I aimed to classify water samples as safe or unsafe based on 10 chemical features. I found that the self-class imbalance was significant, with non-potable samples heavily outweighing potable ones. The random forest model achieved an accuracy of 67.84%, but the recall for the potable class was critically low at 0.38. To improve this, I recommend balancing the training data to enhance the model's reliability. I invite you to ask questions and engage with the visual data I've included."}