Crypto Currency Price Prediction with Machine Learning Using Python

Abstract

We use and study a wide range of  machine learning methods to predict and trade in the daily  crypto currency market. We teach the algorithms to make  daily market predictions based on how the 100  cryptocurrencies with the most market value change in  price. Based on our research, all of the used models are  able to make estimates that are statistically sound, with  the average accuracy of all crypto currencies falling  between 52.9% and 54.1%. When these accurate numbers  are based on the 10% most confident expectations for  each class and day, they go up to somewhere between  57.5% and 59.5%. A well-known case study in the field  of data science looks at how people try to figure out how  much different digital currencies are worth. Stock prices  and the prices of cryptocurrencies are based on more than  just the amount of buy and sell orders. At the moment, the  government's financial policies about digital currencies  affect how the prices of these things change. People's  views about a crypto currency or a star who directly or  indirectly backs a crypto currency can also cause a big  rise in buying and selling of that currency. This study  looks at the trustworthiness of the three most famous  coins on the market today: bitcoin, how well buying  strategies for ethereum and litecoin that are based on  machine learning work. The models are checked and  tested with both good and bad market situations. This lets  us figure out how accurate the forecasts are in light of any  changes in how the market feels between the proof and  test times.