Implementing an ensemble classifier based on a trainable combiner for liver disease diagnosis

The liver is the largest gland in the body that plays a role in digestive activities such as digestion. The entire gastrointestinal tract is collected by encephalitis called portal vein and enters the liver, and all the absorbed material from the gastrointestinal tract passes through the liver. Other liver tasks are trapping toxins and treating them by turning them into harmless substances. Also, this liver is responsible for the metabolism of drugs in the body. Participating in prenatal hemorrhage and resistance to infections, providing rapid energy when needed, storing iron, and other liver functions. The liver cells permanently make various proteins, including albumin, prothrombin, fibrinogen, lipoproteins and heparin. The liver acts as an important buffering system for blood glucose, which means that when glucose levels increase after a single meal to a high concentration, insulin secretion also increases and about 60% of the glucose absorbed from the intestine immediately into glycogen Converted and stored in the liver Machine learning is used in medical diagnostics and has achieved very good results. We used an advanced model to diagnose liver disease in this thesis. The proposed model in this thesis is made up of a seven-class combination. The innovation of this proposed method is in their combination. We use a class to combine seven classes and provide the final answer. In the second chapter, we reviewed the articles presented with the topic of diagnosis of liver disease using artificial intelligence-based methods. By analyzing the articles, we concluded that the combination methods yielded more reliable results. In this thesis, a trained combinatory is used to combine seven grades. According to the last part of the third chapter, we created the new data set. By analyzing the results, we concluded that the combined method of the seven-class stratified individually had a dramatic improvement in efficiency. The result of the proposed combination method is 87/1