Abstract Food analysis is a prerequisite for ascertaining product quality, implementing regulatory enforcements, checking compliance with national and international food standards, contracting specifications and nutrient labeling requirements. One third of all food produced is lost or wasted – around 1.3 billion tonnes of food –costing the global economy close to $940 billion each year. If one quarter of the food currently lost or wasted could be saved, it would be enough to feed 870 million hungry people. In this project, we will be using Machine Learning Algorithms to predict the number of orders a restaurant should produce given certain variables to minimise the food wastage. 7

1. Introduction: 1.1 Overview: Food is central to human well-being: it provides the body with nourishment, offers livelihoods that lift people out of poverty, and brings communities together. Although food is a basic human need, too many people are trapped in a cycle of hunger by forces beyond their immediate control, like poverty, disaster, conflict and inequality. Despite decades of progress in reducing world hunger, 2017 saw increases in the number of people who are hungry. More than 820 million people still go to bed hungry every night — that’s one in every nine people who don’t have the food they need to live a healthy, productive life. The World Health Organization considers this to be the single greatest threat to global health. Hunger is cyclical and generational: it inhibits people’s ability to work and learn to their fullest potential, which can curb their future and trap them and their families in more poverty — and more hunger. People in poverty generally spend between 60 and 80 percent of their income on food, which can force them to prioritize feeding their families over meeting other basic needs or reaching long-term goals, like sending their children to school. If an emergency strikes, they may need to skip meals in order to cope financially — and the cycle of hunger continues. According to the Food Security Information Network, conflict and insecurity were primary drivers of food insecurity in 2017, alone accountable for putting 74 million people in need of urgent assistance. Climate change is also eroding existing efforts to improve food security. 8 1.2 Purpose:. Recent years, the ML methods have become popular as they allow researchers to improve prediction accuracy and used in many applications. In this study, regression algorithms are used to predict various food material requirement. Based on this predictions farmers and producers can produce them to meet the requirement. Here ML regression methods are used for prediction. The study aimed to determine the most successful regression methods by comparing the logistic regression, decision tree (DT), random forest (RF), linear regression, multiple regression, support vector regression, Naïve Bayes. 9
2. Literature Survey: 2.1 Existing Problem: Our world population is expected to grow from 7.3 billion today to 9.7 billion by 2050. Finding solutions for feeding the growing world population has become a hot topic for food and agriculture organizations, entrepreneurs and philanthropists. These solutions range from changing the way we grow our food to changing the way we eat. To make things harder, the world's climate is agriculture. Hence, it is necessary that we analyze the food production and act faster rather than repenting later. 2.2 Proposed Solution : Using a machine learning model, we can predict the number of orders on a food item per day for any kind of cuisine located in various regions and cities which is directly proportional to the food required. We can take various inputs as a base price, region code, city code, food item id, type of cuisine etc to the algorithm. We then use regression algorithms to predict which food item gets how many orders. A food item is considered to be most interesting when the number of orders has more than 500 per day. The model will be trained based on the algorithm features which will give the higher accuracy. Finally, it will be integrated into a web based application. The final system allows the user to give the details required as input and the system will display the output number of orders. 10
3. Theoretical Analysis 3.1 Hardware / Software designing: The software used to implement this project is mainly Jupyter Notebook and Spyder. The other programming skills required are Python, Flask, HTML and CSS. The project works in 2 parts. The first part is building the model and the other part is building the frontend. The model building is done by using Machine Learning libraries by the help of Python. The frontend development is done by using HTML and the designs are added by using CSS. Both the frontend and backend are connected by using Flask. The Flask generates localhost for the application to run. 11 3.2 Block Diagram Food requirement analysis Data Collection Data Preprocessing Import libraries Import data set Data Visualization Handling null values Label Encoding OneHot Encoding Feature Scaling Splitting data Model Building Application Building Create html file Build python code 12 3.3 Flowchart: 13 4.METHODOLOGY: 4.1 DataPreprocessing Data Merging ● Data merging is the process of combining two or more data sets into a single data set. Most often, this process is necessary when you have raw data stored in multiple files, worksheets, or data tables, that you want to analyze all in one go. ● We have three datasets. We merge them to form a single dataset. 14 Data Cleaning ● Data cleaning is one of the important parts of machine learning. It plays a significant part in building a model. Data Cleaning is one of those things that everyone does but no one really talks about. It surely isn’t the fanciest part of machine learning and at the same time, there aren’t any hidden tricks or secrets to uncover. However, proper data cleaning can make or break your project. Professional data scientists usually spend a very large portion of their time on this step. ● Before we go for modeling the data, we have to check whether the data is cleaned or not. And after cleaning, we have to structure the Data. For the cleaning part, First I have to check whether there exists any missing values. For that I am using the code snippet isnull() There are no missing values in our dataset. So we continued with other operations. 15 4.2 LABEL ENCODING: Label Encoding refers to converting the labels into numeric form so as to convert it into the machine-readable form. Machine learning algorithms can then decide in a better way on how those labels must be operated. It is an important preprocessing step for the structured dataset in supervised learning. 16 4.3 ONE HOT ENCODING: A one hot encoding allows the representation of categorical data to be more expressive. Many machine learning algorithms cannot work with categorical data directly. The categories must be converted into numbers. 17 4.4 MODELLING IMPORTANT PYTHON LIBRARIES ● Numpy ● Pandas ● Scikit Learn In modelling the data is fed into machine learning algorithms. So we split the data into two parts training and testing. The training data is used to train the algorithm and testing data is used to test the model . Generally the model having the highest accuracy is chosen as the good model. The problem here we are dealing comes under the classification as we have to predict whether the applicant can get loan sanctioned or not? There are many algorithms for Regression in the scikit learn library. The algorithms here we are going to use are