 Hello everyone. Welcome to the course of business forecasting. Today, we will discuss different type of forecasting methods. In general, in business forecasting or say in predictive analytics, there are different approaches company follow to estimate the demand, sales and financial related instruments. So, today we will understand what are the basic approaches exist in business domain as far as forecasting are concerned, predictive analytics are concerned and then what are the methods we need to study that overall different type of forecasting methods we are going to discuss today. If you look at the general approach, say you know in the introduction session I have already discussed detail of different approaches, but here let us recap them and then we will enter into different forecasting methods or say techniques available in the literature or in the academic then you need to learn. So, before that as I mentioned in the introduction session, in general there are four approaches of forecasting practices that happens in the industry. These are bottom up method, top down method, historical analogy method and deductive method. For example, if you think about say bottom up method, in bottom up method what happens? In any project or any forecasting approach or any particular assignment what the company do is that they ask the unit level managers or say bottom level divisions to estimate the cost, time and say you know say the representative who are related to say close to consumers. So, their opinion they collect and then they estimate from different units from different divisions and from different region and then they collect it after aggregating them the bottom level managers or bottom level units they pass it to the total forecast or total estimate they pass it to the top management and then top management take the final call that is called bottom up approach. Like from the bottom level they estimate different units or the division on the demographic managers they estimate the demand sales forecast whatever prediction you want to do and then they aggregate all them and then pass it to the top management. But in top down method it is different. In this method what they do? They actually the top management take the decision first by a group decision making by inviting experts and they are in a internal executive officer they sit together and they take the decision how much money should be allotted for a project, how much time should be required to completion of a project or what could be the tentative sales demand whether new product need to be launched or not. So, that first it is been decided or predicted or the expected planning are been done in the top level and then that estimation are been passed to the you know bottom level managers or the divisions or the units and then they segregate it to as per the proportionate requirement or the allotment they pass it to the different units. This approach of estimation and the prediction or forecasting are called the top down approach. This is also popular but there are merits, demerits and you know the advantages all this I have discussed in the introductory session. You can refer to that particular lecture. In order to avoid the you know disadvantages like as I mentioned the in bottom up approach what happens? The unit managers they make some additional estimate as a buffer or say safety stock or overhead stock or overhead cost contingency plan they add additional information like you know time cost budget whatever they want to make forecast. Their estimation sometimes are been inflated and when you add all them actually there will be a high like inflated prediction or say you know calculation in that final level of when you aggregate all them. So, there might be a high chance of additional cost additional time which might not be required. As I mentioned because of say risk management and the weather issue technical issue like delay issue they keep additional time additional cost etc. in their forecasting process. In top down method the estimation may be less as compared to the requirement of the actual execution process. So, here you might require quite high cost or say quite additional time or forecasting may be different. Top management might do some over estimated expectation in terms of cost time less cost less time high demand high sale this type of expectation may vary from top management to bottom management. So, in that case you know to overcome that there is a method called you know iterative process or brainstorming process. You can do that also you can follow that like you know brainstorming process in Delphi method we will discuss detail of that today. And then you know there is another process called the historic analogy. In that case the experts rely on the past data and based on that past data behavior pattern and the experience what happened in a particular location or particular project earlier that analogy like you know similar experience similar execution can be done in a new project new forecast new demand planning can be done new product line extension can be done this called historical analogy. You do not spend much time on analyzing the data the similar pattern whatever has happened in the past similar experience similar type of product line or you know new project or new forecast new planning can be done based on the historical experience and historical analogy based on the past data pattern like similar execution you can do in a new project for new sales new prediction etc. But in deductive method what you do we actually do not rely on the data it is a you know opposite of historical analogy in that case the experts feel that whatever has happened in the past that may not reoccur again in a new project new product line or new location or a new prediction in that case do not rely on the past data based on the experience and the current situation the market need we should estimate the prediction the forecast accordingly. So, here based on experience knowledge on the principles of forecasting techniques you can make some prediction based on the experts opinion and do not rely too much on past data and if you can make a forecast based on experience and you know to some extent based on the current contemporary situation trend and the current demand that process of forecasting or approach of analysis are called the deductive method. So, these are the four approaches whether internal issues or internal way of forecasting external input data all this can be put part of all these four approaches, but these are the general common four approaches in forecasting domain that the industry actually follow. So, these four bottom up method top down method historical analogy and deductive method these are the four approaches, but not the technique, but when you adapt or follow any of these four approaches in making forecast or taking a decision in any case application you need some methods and techniques. So, what are the techniques are you know to some extent forecasting tools are available that we need to study as a part of predictive analytics or business forecasting course. So, in general we define the forecasting methods of the techniques into two category one is subjective approach another is the objective approach. In subjective category of forecasting methods or say techniques we follow some qualitative aspects of forecasting estimation like market research consumer survey and then say jury of experts opinion or say you know Delphi method panel of expert group decision making. So, these are called subjective approach here we do not focus much on say quantitative approach of forecasting. There is another approach which is most popular or more famous in the domain of forecasting that is called objective approach under that we prefer or we use different quantitative methods of forecasting. So, what are the quantitative methods of forecasting are available with us it could be time series methods it could be different statistical techniques I am talking about the objective approach mathematical formulation or mathematical techniques. It could be say Arima model as we will be discussing that it could be say you know different causal models of regression analysis simulation models. So, all these you know techniques falls under objective approach or quantity methods of forecasting. Here you will see the general four category of forecasting methods in predictive analytics and machine learning techniques we will cover actually all four method actually all four are important. First one is the qualitative methods which we will discuss today majorly we will concentrate that today and the second is the time series method and then the causal models or we can say it is as a econometric models or regression models and the last one is the simulation models. In this particular course of business forecasting actually we will cover all four today I will concentrate on the qualitative methods majorly what are the qualitative methods and then we will go in the forthcoming session details of time series models and the econometric models like causal models regression analysis and the simulation models also like Monte Carlo simulation system dynamics and risk analytics type of models using simulations all these techniques or methods will be discussed throughout this course different modules of the course. Now if you come to the qualitative forecasting approaches there are actually different approaches are in exist in the market. So, people can follow any type of qualitative approach where you know they feel it is more comfortable right you can have the interview with the managers or you know the experts and then you can you know you can collect the recording of that you know interview process and you can extract that through in view software and you can get a sentiment analysis of this data that also falls on the qualitative aspect also because you know in practice you might not get the data. So, other for new product or suppose MSME sector if you go there they are to some extent you know not that organized. So, in that case if you go and if you you know do some study or research over there what is happening whether they are maintaining any social sustainability or not they are maintaining any environmental you know requirement or not. So, if you take that type of information you might not get the data for that in the industry which has been stored as historical data. So, in that case you have to go so, qualitative interview process you have to conduct and then you have to collect the data from them. So, there are many way you can analyze the data, but in generally as far as business forecasting are concerned we will concentrate into this four category which are more relevant to the course of business forecasting. So, let us go through one by one the first one is the consumer survey or the customer survey or market research. The example that I was talking say if you go to the MSME sector and if you take the managers or the experts opinion that whether they are maintaining managing any environmental you know activities or not or you know social sustainability practices or not you might not get data. So, in that case you are not taking the consumers opinion you are taking the managerial opinion etcetera that can also be a part of your qualitative survey. So, here I have made as a market research or you can say as a like you know company's opinion the experts opinion or you can say the consumer survey or you can put a different terminology also I have made as a market research or say consumer customer survey. So, this what actually the field that actually you are going to the consumer directly or to the experts directly you are going to the field the end customer end person who are practicing the you know activities you are taking their opinion. So, in that case and if you say if you can create a Google seat or you can go to the directly suppose you want to study some FMCG product whether you know say you know pep student is better or say you know say Colgate or say you know any kind of say Ayurvedic pace is better. So, if you want to do a study comparative analysis over there we can go to the different shopping mall and you can ask different consumers or you can create a Google seat or you can take the consumers opinion. We can change the demographic also suppose you can study that in Bombay you can study that in Calcutta you can study that in Delhi or see in Bangalore Chennai and then you can collect the data from different consumers who is consuming the product or the store managers or what kind of sales are happening end to end completely last mile you are reaching and you are taking their opinion and if you do analysis over there you can use quantitative techniques as a subset of that, but overall if you take this type of different for different sources if you collect the data and if you do analysis through you know social media you can collect the data also you can create you can approach the alumni also you can approach different you know consumers and if you collect the data and if you do analysis over there that type of study called as a consumer survey research or market research. In marketing people actually prefer that and they do these practices frequently. So, this is the consumer survey research this can be practiced only through a practical small project or you know practical illustration what in a class it is not possible to do that what you can do is that you can take a theme you can create a Google sheet and you can spread it to your friend circle or to the seniors or the different industry experts or you can say you know directly to the consumers and collect the data and do the analysis over there and you can come up with a forecast planning that what could be the sales feature projects projection of the production or whatever you can do that type of survey or fall under market research or consumer survey. The next level of forecasting approach is that the analysis data analysis approach or estimation approach are called the sales for composite. What is that we call it as a representative sales representative in that case those who actually are in the middle person between the consumers or the market exactly last month consumers or the experts and the company in between the person they are the representative they actually sales the product they know the taste of the consumers or the what the companies or the SMEs or the small store owners are selling the product they know their taste they know what is going on that particular segment. So, therefore, we in that case in second stage like you know in this sales force composite or sales representatives opinion approach there you do not take the direct data from the retailers or directly from the owners or managers from the company or say consumers data you directly take the data from the sales force opinion. So, if you take the opinion from the sales for the representatives who are you know roaming around the areas and dealing the product selling the products or making the marketing of the products you could take their opinion and if you do analysis over there and you can take your expected prediction or the you know whatever the techniques you can use over there and you can make a study over there through these consumers behavior or consumer understanding through or the market test through the sales for representatives opinion that experts opinions are fall under the or the corresponding prediction or fall under the category of sales force composite or you can define any other terminology whichever you feel, but I have given a terminology called as a sales force composite of qualitative forecasting approach. Remember in every where you can use quantitative approaches subset also you collect the data and the analysis you can do it, but the approach overall approach are fall under the qualitative aspects of forecasting or to some extent subjective approach under sales force composite method. Then the third one is very popular in India in everywhere also you know that is called executive committee consensus. Here the company forms a team and then in that team you know some experts the experienced people sit over there some brainstorming sessions happen and on the table say some internal and external experts will sit there some competitor like you know or senior retired person can also come and they can also you know sit over there some academic person also be there who has the experience on that particular field and they the company form a team like you know in your in your college or say you know in industry everywhere in your club everywhere if company or the people think some some problem is occurring they form a team right and that the team or the committee that committee come up with the decision through a brainstorming session maybe one day two day etc and then they come up with the forecast or come up with the decision making that falls under you know to some extent executive committee consensus. This is very popular not only in the forecasting domain it is in the decision making whether the political decision making or you know in the academic level or you know everywhere people form a committee. So therefore in industry also whenever you need to take a decision we call it as a to some extent you know top-down decision making where the top management sits together and they take a decision you can make it as a bottom up also but all aspects can be put under the executive committee consensus where you have a group of people experts and that expert sits together and they take the decision there can be you know they can if required they can take the sales force opinion or sales force persons can be a representative of that committee or even they can approach they can follow the market research approach also. So that could be a subset of executive consensus committee so they will see that what was the market research first what is the opinion from the consumers let us see that and then understand that and then use it in our decision making. So overall executive committee consensus is a superset of first two and where you can take a better decision in terms of group decision making. So there you have a merit that you know you are collecting data from different sources and you are bringing all experts and you have a brainstorming session and then you are taking a decision but there might be a chance that there might be a you know biasness also that means suppose you top persons or the experienced people of the company sits or the supervisor sits over there or the senior manager they are the executive director they may say that no I want this way I want this outcome so therefore what happens you know there may be some sacred cow concept may come into the picture that means whatever the term will say the lower division has to accept that so they are in that case there may be in a deviation from the actual forecast or biasness may come into the picture but it has a merit also because group decision making is being done over there so all people's opinion are being captured and then there is a consensus over there and then they take a decision that is called executive committee consensus which is very popular which also falls under the collaborative forecasting method the last one that is called delphi method which is a superset of you know executive committee consensus and it is actually the most powerful in industry it is not only in the forecasting method it can be used in the other domain also so what is delphi method delphi method is actually iterative process the round of you know like what executive committee consensus take a decision the panel of experts take a decision at one time right but in delphi method you form different you know experts panel so there will not be only one experts panel you can form a two three or four five experts panel and each panels will come up with their opinion let me go to the next slides you will get to know so each panel look at the four figure here you will get to know I will come to that at a later stage let me you know give a introduction of delphi method then I will go to that so in delphi method what happens you prepare four like four five groups of experts in each group there may be one person there may be five person depending on the situation suppose you go for a college competition right so your college goes as a team to compete with some event so other college comes also so this is called different group so in that case all for a single event for a single projects all group will come with their opinion and then there will be a coordinator or the experts team who will who has experience to understand the topic they will take the opinion of each groups and they will evaluate the forecast of each group and here one unique guidelines is there is that the identity of the each of the groups will not be discussed to other group so that means who are the other member you don't know they are they will come they will give their feedback you will come you will give your give your forecast and overall the coordinator will see and they will give a feedback to each group and the each group will come up in the say after seven days of the next round they will come up with their you know revised forecast and then there will be at a at a later stage after couple of iteration of revision revising the forecast of each groups at a later stage you will see there will be a consensus like a pyramidical format so so there will be a final consensus of each groups i'll discuss that more detail in the forthcoming slide you'll get to know here I have put a example of all four then I'll go to the detail of Delphi method with the example so here you can see for consumers market research here you can see I have taken an example of you know e-commerce domain as a part of example and here you can see you conduct the survey and take the opinion whether it could be you know like websites you know product selection consumers service customer service delivery speeds whatever you know topic you want to like online shopping experience whatever the topic you want to capture not a matter you can take the feedback of the consumers and you can take a decision of that by analyzing that this is called consumer market survey here with the example with the same domain e-commerce field I have kept for your quick understanding here you can see you take the opinion of the sales force team of the e-commerce industry or the company and if you take the experts opinion who are the sales force who are closer to the like you know closer to the consumer side if you take their opinion if you take a decision that will fall under sales force composite method now you think about third one the say jury of experts opinion here you can see the in the brainstorming session the experts panel are being formed and the experts will all the experts from the different division as the example I was giving so from the different division different experts experience people will sit and they throw a brainstorming session they will come up with a forecast that forecast will fall under jury of experts opinion what about the field you can take I have given example of e-commerce platform you can take any other area also not a matter same area you can take or I mean same approach you can follow now look at the delphi method here what do you have here you can see here we have formed a experts only one person I have put it does not mean that there will be one experts you can have a more more member also suppose here is one team here is another team say here is another team here is another team here is another team so suppose four groups we have formed so four group will come up with their opinion and then they will pass their feedback their forecast to the experts to the panel coordinator and the panel coordinator will evaluate their opinion and he will give a feedback like this he will give a feedback suppose it is round one it is round one so all say groups suppose four groups four group will group one group two say group three say group four right all group will come up with their forecast their opinion in round one and then in round two in round two suppose the coordinator will give a feedback to each of the group member because you know each group will not understand will not know the identity of the other group so that will not be disclosed so therefore only coordinator know who is participating in the forecasting process from a particular group and then each group will come up with their forecast with the revised forecast at a later stage after couple of round you will see the consensus initially there will be you know deviation there will be diversion in the third process but at a later stage there will be convergence this process is very crucial because you are changing every group's opinion through feedback process and there will be consensus without disclosing you know the identity of the other groups what they are project projecting this process is very strong and we call it as a Delphi method it is not only popular in industry in terms of forecasting process it is popular in other group decision making also here is one example of Delphi I have kept for your quick understanding. Suppose you want to make EV projection in India by 2030 you use the Delphi method remember round one, round two, round three suppose and and then you will take a final decision so as I was mentioning suppose here suppose you have say one group two group three group four group five group and to estimate the EV sales in India by 2030 suppose for the sake of illustration only just randomly I have prepared it so suppose you will take the experts opinion in each group you have formed a group experts groups suppose you have a five group here one two say one two three four five suppose five group right so suppose you have formed a five group so each group will come up with round one each group will come up with their forecast suppose here 10 million 15 millions per year sales by 2030 and say 20 millions 10 millions 10 millions they have come up with their forecast there might be mean standard deviation etc all the analysis they can do they can use quantity approach here also in their data collection process and analysis also they can use experts opinion they can have a group decision making they can have a consumer survey market research all this they can do and they can come up with say secondary data analysis tertiary data like you know document analysis so they can do all these things after that they will come up with their forecast this forecast will be passed to the coordinator right will be passed to the coordinator say suppose some coordinator will be there suppose you know some coordinator will be there so coordinator will take the decision about the decision making of each group or what what forecast they are giving so they will give a feedback the coordinator and the team of experts will give a feedback to the or you know the final decision maker team will give a feedback to them there then you might ask that if there is a coordinator the final decision maker then we can form a team for them also no they only they have the experience only they will not take any decision they will rely on the decisions of the experts panel but they will evaluate all of them how much deviations are they are in from the group from one group to other group why this group are coming up too much of diverse like with different opinion suppose we did a study during our corona time I did a study in my college that students were at home and they were discussing that college were discussing that my institutes that whether you know when the offline class will start so I thought of you know doing a Delphi process over there and I asked the all the students they form a group of 10 members of each group and then you come up with their predictions that when we can open our you know offline classes in the campus so every group come up with their prediction in the first week so somebody says that maybe you know end of 2020 2020 December we can like you know from January of 2021 we can start somebody said I remember that you know some group that we will start we cannot start before July 2021 somebody says that by March we can come by 2021 March we can come back to the campus and the classes can be offline itself like in person itself so that was a different opinion so we did that study in that in 2020 around 2020 August or September so we did a study but you can see the prediction some group says that we are saying that by end of December 2020 the classes can start offline classes in the campus some group we are saying that by April or May March the classes can be started some by next year 2021 some group are saying we are saying that no we cannot start by like July or August of the next year even some group said that we cannot start before 2022 that time they said so this is what the different opinion then I have given opinion like you know feedback to all of them why you are dividing deviated from them somebody is saying that we can start in the 2020 December itself and we are saying that by 2022 January we will start the class before that we cannot start the class what is the opinion from what does the you know support for your decision making can you come up with the revised forecast so everybody come up with revised forecast in second round then I have given a feedback further so we have to see the I have not I am not taking any decision so coordinator will not take any decision just coordinator will give the feedback to each group's opinion that why you are deviated from others so this way what happens you know we have come up with the opinion that was of for that particular case illustration of you know when to start the offline classes there it was around you know the final consensus for all the groups were around end of 2021 but effectively you know we started around 2022 actually so that is a different part end of 2021 actually we have started so that is a different part but here you can see the example here say each group have come up with their forecast how many AV vehicle sales will be there in India by 2030 say so each group have come up with their opinion then I have given a feedback suppose coordinator can give a feedback to them and then you can see the consensus to some extent deviation has come back so in round 2 what happens you know this this deviation will come back and over a period of time you will see the consensus so suppose you will see the consensus over a period of maybe after two round three round etc to some extent and that could be considered as a forecast here the qualitative approach of forecasting that means this one we are actually jury of experts opinion we have taken only one group opinion just one group opinion in three round we have done a three round of revision for one group opinion and we have given feedback to other group also so you can see the covariance among the data sets among the groups and you can see a better forecast with a better accuracy with a better convergence this process called is called as a Delphi method and it is very popular in the forecasting techniques and even industrial decision making also so these are the four methods of you know qualitative approach there are many methods many other methods also we can discuss but let us restrict the discussion of these four methods we will try to you know like illustrate one or two approach of qualitative forecasting that we have discussed today this four approach that we have discussed we will try to you know illustrate one or two through assignment and you can come up with their forecast planning and we can have a Delphi approach over there we will see that time what could happen what could be the forecast for a particular project or particular topic we will illustrate that now let us go to the another approach of forecasting that is called quantity forecasting I just told that there are majorly three approach of quantity forecasting one is a time series methods time series analysis where we will spend maximum of our time and then there will be a regression analysis or causal models or econometric models we will spend significant amount of time for them also and then like simple linear regression multiple regression logistic regression as a part of method learning approaches we will discuss that also then we will go to little module on Monte Carlo simulation also how simulation approaches are are been like you know defined through which how through which how you can make a forecast for you know for practical problem where you have a complex system and what is the like you know mechanism of Monte Carlo simulation or system dynamics approach and how you can handle risk modeling through simulations this type of different models we will study through simulation approach of forecasting also now come to the time series analysis time series analysis has the unique approach that that data will be collected over a period of time and here you will study the behavior of a time series data based on its past performance so whatever the data you have so this is your behavior I passed behavior based on that you want to study the future of same data look at the whatever the data you have that is your past behavior may be you know stock price it may be you know whether it may be rainfall it may be you know crude oil price it may be you know population anything and company earning you have the data so whatever has happened in the past you take that as a as a support of your data and study the behavior of your own that is all dependent variable there is no independent for only one set of variables and you want to study the behavior of that for the future that is called time series data analysis it can be daily data it can be monthly data it can be quarterly data it can be yearly data also in the type of quality type of forecasting methods we are discussing this much only in detail we will study the various models of time series techniques like moving average techniques like you know exponential smoothing metals models like you know hold model winter model arima model air process MA process then we will discuss decomposition method many methods will be discussed throughout this courses of business forecasting and entire concepts of time series analysis will be done through examples through acceleration and through you know different technical understanding also then you can see the last statement also the goal of time series analysis is to extract the meaningful information from the past data extract the meaningful information from the past data and make a future forecast detect the anomalies if you have you know identify the pattern in the data like pattern in the data so, whether it is a seasonality is involved or not whether it is uptrend data or not some randomness information is involved over there or not some you know stationarity is involved or stationarity or non-stationarity is involved in the data or not you study the behavior and then make the forecast for the future this is called the time series data analysis then the second part among the out of quantitative techniques are called the causal models or regression models this is very popular here the difference between the time series analysis and regression model are very unique that is in time series you have only independent variable in time series data you have only dependent variable you do not have any independent variable but in causal models or regression in regression analysis you have dependent variable you have independent variable so, maybe more than one independent variable suppose your work experience and your academic performance both will give you your performance in the company or job opportunity in the company for the for a placement so, here you can have a say you know different candidates data you can collect you know so, like this way you can get the data of each candidates from the past data sets or say suppose you want to give a loan to a person so, whether the persons how much loan the person should get you can take the decision based on the person's salary person's civil score person's asset etcetera so, these are the past different candidates data you can gather as a independent variables and the performance whether how much loan has been disbursed to a person in the in the previous data sets that will be your dependent variable so, this way you will be having independent variable data sets and the dependent variable data sets and you found the causal relationship causal effect relationship you found the explained relationship between them that is called the regression analysis how you are regressing the relationship between the dependent variable and independent variable that is called the regression analysis fit the look at the here some data points are being posted here and here time is not a matter just sample of data is important for this causal relationship causal models or regression analysis Suppose, one independent variable say X, you have collected different data of Y and you have put in the graph and you are trying to find the best relationship among them say linear regression you are finding. So, this is called regression analysis or causal relationship between dependent variable and independent variable. That means, how much the dependent variable is explained by the independent variable that is called regression analysis. We will study details of regression analysis through simple linear regression through multiple linear regression, multicollinearity concepts, we will also study the logistic regression where you can take a binary or categorical decision making which is very popular in industrial. So, like I was talking about if you want to give a loan to a person whether you should give a loan or not that is the first part, then how much loan you want to give, whether if somebody come up with a insurance claim, whether the insurance claim is a fraud claim or it is a true claim. So, that is a decision making. So, you have to take a binary decision variable where the outcome will be 0 or yes or no. So, how you can take a decision that also falls under logistic regression study that also under the context of regression analysis or you know causal method or we call to some extent we all we can define as a part of predictive analytics or say machine learning approaches. The last one is simulation models here generally we imitate the process for a complex system by giving input data of variables or parameters and then we take a decision of the variables and we imitate the process for thousands of iteration. In a second of through computer simulation through computer simulation within a second we can generate thousands of output of this complex system. So, here you have a complex system say and here you give the input of the parameters and through by generating random number we will discuss detail of it and you imitate the output of output this is input data you are giving in every simulation every iteration and you simulate the process and you get the output of the decision of the system and you generate these data sets for a given input what is their output of the system. So, these data sets you can record thousands of data through computer simulation within a second or within a minute and then you can calculate the prediction of the behavior of the system and we can make a better prediction over there. The expected outcome the confidence interval the correlation between the data and you know that the scatter plot all this you can draw and you can take a better decision through Monte Carlo simulation or system behavior system dynamics study we will spend some time. So, two three sessions or two three modules over there as a part of simulation models and the corresponding forecasting risk analytics can also be done through this process also. So, this is the overall process that you know four models that we can we will discuss one is the quality forecasting which we have discussed today and the time series models there are many time series models I have listed here all of them will be studied through this course then the causal models regression analysis all this regression models will also be studied through this course a simple multiple and the logistic regression and the simulation study will be also covered Monte Carlo simulation system dynamics approach the introduction of detail of Monte Carlo simulation basic concept of system dynamics approach continuous version of simulation and the risk modeling using simulation through different practical case studies with say you know insurance industry supply chain analytics or even say product mix production planning problem or you know portfolio analysis we study through different case approaches. Let us conclude today's session of types of forecasting techniques under the context of business forecasting. Thank you.