| dc.description.abstract | One of the most groundbreaking developments in the financial sector is financial technology (Fin Tech), which is expanding quickly. The market's thirst for alternative financing increased as a result of decreased trust in financial service providers. Fintech is heavily influenced by a number of technological developments, including the accessibility and affordability of infrastructure (such as the Internet, cellular technology, sensors, and increasingly sophisticated technology platforms, Big Data analysis), as well as business practices (such as the sharing economy). Many existing processes of traditional financing and investment are taking a simplified and faster digital approach. The solutions provided by fintech are ground-breaking because they combine numerous financial procedures with the power of technology. P2P lending is one of these FinTech subsystems. It is a more straightforward and quick approach for lenders, borrowers, and the lending platform to work together to make loans. P2P lending offers profitable interest rates for the participating stakeholders by cutting out the typical banking intermediaries, and this streamlined process is acquiring a lot of traction, leading to significant and steady industry growth. The typical checks and balances used by financial institutions cannot be used in P2P lending because it is an evolving and rather distinct business model. The risk of investing in P2P lending is the most difficult aspect of P2P lending since it threatens the investors who play a crucial role in the entire industry. The failure to correctly estimate the borrower's default is the main factor driving risk. For this project, we analyzed P2P lending markets and wanted to address the imperative factor of forecasting delinquent customers that was inhibiting investors from entering this market. In this process we also utilized many ML algorithms to come up with the classifier model that has high accuracy. Many P2P lending platform datasets were considered for this study and finally Bondora FI dataset was picked because of its richness in variety of features and extensive datapoints. The research and analysis followed a 2 phased methodology. The first phase was focused on the data and the second phase on the model. Using multiple algorithms, we came across the best model result from the Random Forest algorithm. This project also covers the importance of identifying the significance of the given dimensions in the dataset and further assisting the P2P lending platforms in implementing their accurate classifier model depending on their borrower’s pool. P2P lending platforms can utilize the model to determine the potential delinquent borrowers and further mitigate the risks by altering interest rates or restructuring payments. Firms can establish their strategies by providing data driven insights to their investors with the help of an AI based accurate model. The paper also derives the massive business insights that the firms in their operations can utilize. The AI driven futuristic outlook stemmed from the analysis focuses on the 3 critical phases of the value-based strategy – Value creation, value delivery, value capture. The intelligence delivered from this research will not only impact the P2P platforms to control their default rates but also enhance the trust factors among the investors and attract more parties into the platform with a perfect measure of the interest rates involving minimal risk. Lastly, this paper talks about the future potential research pertaining to the areas like data collection, model building, and futuristic business use cases that can be derived in P2P lending markets. | en_US |
