a logistic regression unit, which is a single neuron neural network. Implement it using Python code in a Jupyter notebook. Use the Log Loss Function and write code for stochastic gradient descent (SGD) to optimize the network. Additionally, replace the sigmoid activation function with ReLU and tan functions. Organize your implementation as a set of functions*.

Apply the logistic regression unit on a classification problem using a dataset obtained from the internet, such as the housing dataset. You can use the Kaggle housing datasets available at any of the following links:

- [Kaggle Housing Prices Competition - Advanced Regression Techniques] (https://www.kaggle.com/c/house-prices-advanced-regression-techniques/dataLinks to an external site.)

- [California Housing Prices] (https://www.kaggle.com/harrywang/housingLinks to an external site.)

- [Ames Housing Dataset] (https://www.kaggle.com/c/house-prices-advanced-regression-techniquesLinks to an external site.)

Evaluate the performance of the model by calculating the accuracy and F1 score of the results. Write Python code in the Jupyter notebook for this task.

*Note: You are not allowed to utilize machine learning packages or libraries like scikit-learn for implementing this algorithm. Using each package will result in a deduction of 5 points. However, you are allowed to use pandas, NumPy, matplotlib, seaborn, and scipy.