Embedded Files
Goal: Aim of this research is to identify crucial elements that contribute to a startup company's success and provide insights to Capitalist for their better investment on emerging startups.
Abstract: The success of a business frequently relies on the early angel investors who take a big risk by believing in and supporting the ideas. Along with them, big companies choose the startup to merge with them or to buy, wealthy investors pick the ideal startup, or even the general public invests in stocks, or the government can use it to help fund companies that are beneficial to society. Predictions can only be based on past data, with little regard for market stability. Hence, it is crucial to evaluate the ideas' quality and amount of creativity and to make a judgment on their likelihood of success or failure.
Approach
Process
Process
Pre-Processing - Performed One-hot encoding for categorical data. Removed NaN, unmatched column data types.
Visualization - This process has been more helpful in understanding the relation between columns, grouping the like columns. As displayed this Correlation displays the top features that are correlated.
Outliers Removal - Used two techniques, Z-score and Interquartile range. Which uses median and Standard deviation to remove the outliers.
Model Prediction & Evaluation
First, dividing the dataset into, 80% training and 20% testing data. The models used are Logistic Regression, Random Forest, Decision Tree.
Used K-fold Cross-validation method to compare the models, in order to find the better model.
To conclude, which model better fit to the dataset. Drafted a graph to calculate ROC curve, which leads to finding the AUC(Area under the curve), comparatively Decision Tree gave better result of 78%.
Poster
Google Sites
Report abuse