Tasks---The-Sparks-Foundation

All the .ipynb files from the Coursera courses i have successfully completed.

PCOS = Polycystic Ovary Syndrome. PCOS is one of the emerging diseases that is seen almost in every woman today due to our lifestyle, stress, anxiety, and malfunctioned food products. It is a complex condition where ovaries produce out of control androgens, i.e., male sex hormones which is present in women in usually small amount with no specific treatment or medication. When a woman is diagnosed with PCOS, it means that numerous small cysts have developed in the ovary hence disrupting normal ovary functions. These days, women when diagnosed with PCOS, they do not share it with their family and it's a taboo disease that comes with a tag that a women cannot get pregnant when she is diagnosed with PCOS. PCOS is never discussed a lot and there is no proper medication to cure this condition. This condition results into high mood fluctuations, intense stress, weight gain, hair loss, acne, disrupted reproductive system, anxiety, skin patches, thyroid etc. I am specifically choosing this dataset as this condition is more serious than it is described, and it cannot be diagnosed at early stage as of now. It becomes very important to diagnose PCOS at early stage to avoid the misconceptions about the conditions and for the proper treatment. PCOS are of 4 types: 1) Insulin Resistant - High level on insulin drives up androgen levels and can develop Diabetes 2) Adrenal PCOS - Excess Adrenal hormones causing infertility, most difficult to be diagnosed 3) Inflammatory PCOS - Excess testosterone resulting in ovulation issues 4) After pill PCOS - This occurs in some people after they stop taking oral contraceptive pills Feature selection becomes the most crucial part here as there are thousands of features available in the dataset and no one of them guarantees the successful diagnosis of PCOS. Here I will be using 4 different Machine learning algorithms: 1)Decision Tree 2)Logistic Regression 3)KNN 4)Random Forest Hence my reason for choosing this dataset is to further extend my analysis by diagnosing the type of PCOS a patient is suffering from, hence avoiding future complications. Currently, as per - the analysis is done only to detect PCOS. The next target would be to determine the type of PCOS from its symptoms to help patient take precautions accordingly. Part B - Obtain Data To proceed with our dataset, data needs to go through couple of steps here. The data is downloaded from Kaggle, and the data is collected from 10 different hospitals in India. The data downloaded is already in the CSV (Comma separated value format). I am working on Anaconda Jupyter notebook. I downloaded Data from Kaggle which was already in .csv format and uploaded it to Jupyter notebook and assigned to 'data' variable to read CSV file in data frame. I have already cleaned my data in excel and dropped columns such as - Sr no, Patient file, BMI, Hip, Waist, Hip:Waist Ratio. I cleaned up the extra space or any null values in excel itself.